Process research and scale-up for a β-3 adrenergic receptor agonist

Article abstract of DOI:10.1021/op049969o

A scaleable synthesis of the β-3 adrenergic receptor agonist, (4-{2-[2-(6-aminopyridin-3-yl)-(2R)-hydroxy-ethylamino]ethoxy}-phenyl)acetic acid is presented. The key coupling step utilizes a silyl-protected β-hydroxy tosylate to alkylate a primary amine,

Full text of DOI:10.1021/op049969o

Organic Process Research & Development 2004, 8, 583−586
Process Research and Scale-Up for a â-3 Adrenergic Receptor Agonist
Brian C. Vanderplas,* Keith M. DeVries,1 Darrell E. Fox, Jeffrey W. Raggon, Michael W. Snyder, and Frank J. Urban
Pfizer Chemical Research and DeVelopment, Groton, Connecticut 06340, U.S.A.
Abstract:
out to be problematic. As reaction size and filtration time
increased, significant racemization of epoxide 6 was seen.
This was presumably due to deprotonation of the amide
nitrogen, leading to equilibrium with the quinone methide
6a (Scheme 2).
Using less than one equivalent of base and keeping the
filtration cold helped to reduce this racemization, but the
process was not robust. Epoxide 6 of varying chiral purity,
ranging from 96% to 40% ee was isolated. A weaker base
such as potassium carbonate in acetone initially performed
well on small scale in the lab without racemization but gave
inconsistent chemical yields on larger scale. Because of these
problems, the epoxide formation step was done in multiple
22-L batches with potasium tert-butoxide in THF to minimize
loss of the expensive diol 4.
The carboxylic acid group in amine 7 was protected as
the N-methylamide. When an alkyl ester protecting group
was tried, the primary amino group in 7 would displace the
alcohol and form an amide. This resulted in oligomeric side
products. More hindered amides such as the N,N-dimethyl-
amide of 7 were more difficult to hydrolyze at the end of
the synthesis. The opening of epoxide 6 with the amino side
chain 7 was initially done in 90 °C dimethyl sulfoxide
(DMSO). This gave poor yields, and the epoxide was found
to be unstable in DMSO at that temperature. By using a 5:1
mixture of toluene and DMSO at 95 °C, epoxide 6 itself
was much more stable, and reaction of 6 with excess 7 in
this solvent system was used. Treatment of the crude reaction
mixture with di-tert-butyl dicarbonate yielded organic soluble
BOC derivatives 10 and, to a lesser extent 11, from which
the water-soluble bis-alkylated products could be washed
away along with DMSO and excess of amine 7. The minor
amount of the isomeric BOC derivative 11 detected suggested
that epoxide opening at the R-carbon to yield alcohol 9 might
not be an issue. However, a mass balance analysis of the
reaction mixture before the protection sequence showed it
to consist of desired 8 (48%), R-opened product 9 (20%),
bis-alkylated 7 (9%), and unreacted 6 (7%). The remaining
components were not identified. At this point, it was realized
that the low levels of 11 observed were due to the slower
rate of reaction of 9 with di-tert-butyl dicarbonate and not
to the lack of formation of 9 itself.
A scaleable synthesis of the â-3 adrenergic receptor agonist,
(4-{2-[2-(6-aminopyridin-3-yl)-(2R)-hydroxy-ethylamino]ethoxy}-
phenyl)acetic acid is presented. The key coupling step utilizes
a silyl-protected â-hydroxy tosylate to alkylate a primary amine,
thus avoiding water-soluble intermediates and reducing un-
wanted side products. Deprotection strategies and isolation of
a zwitterionic species are described.
(4-{2-[2-(6-Amino-pyridin-3-yl)-(2R)-hydroxy-ethylamino]-
ethoxy}phenyl)acetic acid 1 is a â-3 adrenergic receptor
agonist discovered at Pfizer Global Research and Develop-
ment by Robert Dow.2 In this contribution, the efforts to
provide kilogram quantities of 1 are described. The key issues
were selection of protecting groups, the high water solubility
of many of the intermediates, purification of complex
reaction mixtures, and racemization of the single asymmetric
center. The final compound was isolated as the zwitterion.
The original synthesis of 1 was sufficient for the prepara-
tion of multihundred-gram lots of material (Scheme 1).
2-Amino-5-bromopyridine was acetylated and treated under
Heck reaction conditions to provide 2-acetamido-5-vinylpy-
ridine as described by Raggon.3 N-Acyl protecting groups
such as benzoyl were tried but were difficult to hydrolyze,
and imide protecting groups were found to react with primary
amine 74,5 in the coupling step. Introduction of the single
chiral center was accomplished via asymmetric dihydroxy-
lation of olefin 3 to give diol 4. Tosylation of 4 was done
with p-toluenesulfonyl chloride in pyridine at 0 °C to provide
the primary tosylate 5 selectively. Treatment of tosylate 5
with base provided epoxide 6.
Initially, on laboratory scale, sodium hydroxide supported
on alumina was used for this transformation to allow workup
without using water. This was changed to potassium tert-
butoxide in tetrahydrofuran (THF) for further scale-up. Either
of these procedures required a filtration to remove salts
from the water-soluble epoxide 6. These filtrations turned
Refluxing crude isolated BOC-alcohol 10 with 6 N HCl
removed all the protecting groups. Chiral zwitterion 1 was
isolated in 10-15% yield (from epoxide 6) with good chiral
purity following adjustment of the solution to pH 7. While
the throughput was low, the synthesis did allow the prepara-
tion of multihundred-gram quantities of 1 for early studies.
* To
brian_c_vanderplas@groton.pfizer.com.
(1) Current address: Eli Lilly & Co., Indianapolis IN.
(2) Dow, R. L. U.S. Patent 6,001,856 1998; Chem. Abstr. 1999, 130, 81419.
(3) Raggon, J. W.; Snyder, M. W. Org. Process Res. DeV. 2002, 6, 67-69.
(4) DeVries, K. M.; Dow, R. L.; Wright, S. W. WO98/21184.
(5) DeVries, K. M.; Raggon, J. W.; Shanker, R. M.; Urban, F. J.; Vanderplas,
B. C. U.S. Patent 6,124,457, 2000.
whom
correspondence
should
be
addressed.
E-mail:
10.1021/op049969o CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/28/2004
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
583

Scheme 1
Scheme 2
Scheme 3
Efforts to improve this chemistry focused on avoiding the
complex reaction mixture formed in the epoxide-opening step
while looking to improve the organic solubility of the
intermediates (Scheme 3). Crystalline tosylate 5 was pro-
tected as a tert-butyldimethylsilyl ether 12 by treatment with
tert-butyldimethylsilyl chloride and imidazole in N,N-di-
methylformamide.6 The reaction of tosylate 12 with two
equivalents of amine 7 and one equivalent of diisopropyl-
ethylamine was carried out in very concentrated DMSO
solution (0.6 vols with respect to 12) at 80 °C. The reaction
mixture was initially a very thick slurry, which collapsed
into a stirable oil at ca. 50 °C. More dilute reaction
conditions, use of less than two equivalents of 7, or omission
(6) The protection of chlorohydrins as triethylsilyl ethers followed by conversion
to the iodo compound and reaction with a primary amine has been
described: Corey, E. J.; Link, J. O. J. Org. Chem. 1991, 56, 442-447.
Brodfuehrer, P. R.; Smith, P.; Dillon, J. L.; Vemishetti, P. Org. Process
Res. DeV. 1997, 2, 176-178.
of the additional tertiary amine led to incomplete consump-
tion of 12. Higher reaction temperatures induced acetyl
584
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development

transfer from the acetamide group of the pyridine to primary
amine 7. Use of other tertiary amines such as pyridine,
DABCO or N-methylmorpholine resulted in incomplete reac-
tion, and pyridine itself was quaternized by 12. After heating
overnight, the reaction mixture was cooled and diluted with
ethyl acetate. Excess amines were washed away with water,
and the organic layer was concentrated to an oil that consisted
mainly of desired 13 with <2% of bis-alkylated product.
In analogy to the previous route using BOC-derivative
10, a complete deprotection of 13 was carried out by heating
in 6 N HCl. Chiral HPLC analysis of the reaction mixture
showed nearly racemic 1. A variety of desilylation procedures
such as TBAF/THF, TFA/CH2Cl2, HF‚pyridine, and HOAc/
THF/H2O were used to remove the TBDMS group first.
Hydrolysis of the resulting amino alcohol 8 by 6 N HCl gave
racemic 1. In comparison, the methyl ester of 1 (prepared
by treatment of 1 with thionyl chloride in methanol) could
be hydrolyzed by heating in 6 N HCl to afford a high yield
of 1 without measurable racemization. Furthermore, zwit-
terionic 1 could be heated in aqueous HCl and recovered
after pH adjustment without racemization. These observations
suggested that it was the presence of an N-acetamide moiety
in 8 that led to racemization of the benzyl alcohol group.
On the other hand, in the unprotected 2-aminopyridines 1
and its methyl ester, the 2-amino group would be protonated
in 6 N HCl, and this would deactivate the benzylic position
to ionization and racemization. It was also found that only
the excess of desired enantiomer 1 precipitated as the
zwitterion 1 at pH 7, while the material left in aqueous
solution was racemic. This helped explain the low recoveries
seen with the epoxide process.
Two different methods were then developed to convert
13 to 1 in high yield and optical purity. In the simplest
procedure, basic hydrolysis of 13 with excess sodium
hydroxide at reflux removed all the protecting groups. The
cooled reaction mixture was a slurry. Activated charcoal was
added, the mixture was filtered, and the solids were washed
with water. Zwitterion 1 was precipitated by lowering the
pH of the solution to pH 7. While eliminating additional
isolation steps was attractive from a processing prospective,
the removal of the silyl protection in this manner provided
a product contaminated with residual silicon-containing
impurities. The preferred deprotection method was a stepwise
removal of protecting groups. Intermediate 13 was treated
with tetrabutylammonium fluoride in THF to remove the tert-
butyldimethylsilyl moiety, and ethanolic HCl was added to
the reaction mixture to precipitate the dihydrochloride salt
of 8. This isolation of a crystalline solid provided a
convenient purification point. Hydrolysis of the amide
protecting groups in 8 was carried out with aqueous sodium
hydroxide at reflux, and zwitterion 1 was isolated by
adjusting the pH of the solution. Further purification, if
needed, was accomplished by dissolution of the zwitterionic
1, filtration of the solution to remove insolubles, and
reprecipitation of zwitterion 1 by adjustment of the pH of
the aqueous solution to pH 7.
fully dealt with the issue of complex mixtures of water-
soluble compounds. This process was used to provide batches
as large as 30 kg of material suitable for clinical supplies.
Experimental Section
Melting points were determined on a Thomas-Hoover
capillary melting point apparatus and were uncorrected. NMR
spectra were obtained on either a Brucker WM 300 (300
MHz) or a Varian Unity 400 (400 MHz) spectrometer in
deuteriochloroform or DMSO-d6. Infrared spectra were taken
in KBr by diffuse reflectance infrared Fourier transform
spectroscopy (DRIFTS). Mass spectra were determined with
a Finnigan 4510 mass spectrometer using fast atom bom-
bardment (FAB). Elemental analyses were performed by
Schwarzkopf Microanalytical Laboratory, Woodside, N. Y.
Toluene-4-sulfonic Acid 2-(6-Acetylamino-pyridin-3-
yl)-(2R)-(hydroxy)ethyl Ester (5). Chiral diol 4 (71.3 kg,
363.4 mol) was dissolved in pyridine (55 gal) and cooled to
-10 °C. p-Toluenesulfonyl chloride (86.6 kg, 545.3 mol)
was added in three portions over 1.5 h. The reaction mixture
was stirred for 4 h at 0 °C, at which time TLC (9.5:0.5
methylene chloride/methanol) and HPLC analysis indicted
the reaction was complete. Water (34 gal) was added to the
reaction over 60 min. The quenched reaction was added over
20 min to a biphasic mixture of ethyl acetate (157 gal) and
a solution of sodium carbonate (96.3 kg) in water (254 gal),
keeping the reaction temperature between 5 and 10 °C. This
was stirred for 20 min, after which time agitation was
stopped, and the layers were allowed to separate. The organic
layer was washed with water (4 × 63 gal) and concentrated
under vacuum to an oil. Isopropyl alcohol (16.6 gal) and
toluene (168 gal) were added to the oil, and the mixture
stirred for 16 h at 25 °C. Additional toluene (82 gal) was
charged, and the slurry was cooled and stirred at 10 °C for
3 h. The product was collected by filtration, washed with
toluene (15 gal), and dried under vacuum at 50 °C to yield
73.7 kg (57.6%) of 5 as a pale yellow solid; mp 118-122
1
°C. [R]D -48.9° (c ) 1.01, MeOH). HMR (400 MHz,
DMSO-d6) δ 10.42 (s, 1), 8.16 (s, 1), 7.93 (d, 1, 8.7 Hz),
7.7-7.59 (m, 3), 7.37 (d, 2), 4.93 (t, 1), 4.00 (s, 2), 3.28 (s,
1), 2.35 (s, 3), 2.03 (s, 3), 0.73 (s, 9), -0.04 (s, 3), -0.19
(s, 3).
Toluene-4-sulfonic Acid 2-(6-Acetylamino-pyridin-3-
yl)-(2R)-(tert-dimethyl-silyloxy)ethyl Ester (12). Mono-
tosylate 5 (73.7 kg, 210.3 mol) was dissolved in N,N-
dimethylformamide (20 gal) to which were added without
cooling imidazole (28.6 kg, 420.1 mol) and tert-butyldim-
ethylchlorosilane (34.9 kg, 231.5 mol). The reaction mixture
was stirred at 25 °C for 2 h, at which time HPLC analysis
indicated the reaction was incomplete. Additional tert-
butyldimethylchlorosilane (5.50 kg, 36.5 mol) was added,
and the mixture stirred for 12 h at 25 °C, when HPLC
analysis indicated the reaction was complete. The reaction
mixture was added to ethyl acetate (156 gal) and water (78
gal) and stirred for 15 min. Agitation was stopped, and the
layers were allowed to separate. The organic layer was
washed with water (78 gal) and concentrated under vacuum
at 25 °C to approximately 5 gal. Hexanes (78 gal) were
added, and the resulting slurry was cooled to 10 °C and
In summary, a relatively simple process for a large-scale
preparation of amino alcohol 1 was developed that success-
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
585

stirred for 3 h. The product was filtered, washed with hexanes
(15 gal), and dried under vacuum at 50 °C, providing 92.7
kg (94.6%) of 12 as an off-white solid; mp 121-124 °C.
Chiral HPLC7 analysis showed the solid to be 99.67% desired
G-60 (7.4 kg) was added to the reaction mixture, and the
temperature was increased to 65 °C. After stirring slowly
for 30 min, the mixture was filtered warm (55 °C) and the
filter cake washed with warm water (10 gal). The filtrate
was then cooled to 10 °C and adjusted to pH 7 by adding
approximately 9 gal of concentrated HCl while keeping the
temperature between 5 and 15 °C. The resulting slurry was
then stirred at 10 °C for 2 h, filtered, washed with water (55
gal), and dried under vacuum at 50 °C for 16 h, yielding
108.5 kg crude zwitterion 1 as a water-wet solid. LOD
analysis showed this product to contain 54% water; thus,
the actual yield was 49.9 kg (92.8%). For analytical purposes,
a small sample was dissolved in dilute HCl, the insolubles
were filtered off, and 1 was precipitated at pH 7 by adding
dilute aqueous NaOH. Filtering and drying under vacuum
at 50 °C furnished an off-white solid, mp 210-211 °C.
To a mixture of crude zwitterion 1 (49.9 kg, 150.6 mol)
in water (53 gal) was added concentrated HCl (6.57 gal) over
30 min, keeping the reaction temperature between 20 and
25 °C. The hazy mixture (pH approximately 0.8) was stirred
for 10 min at 25 °C, then filtered through a pad of Celite.
The pH of the filtrate was adjusted to 7.0-7.2 with 10%
aqueous NaOH8 (prepared by dissolving NaOH flakes (12.05
kg) in water (31 gal)). The resulting slurry was filtered and
washed with water (26 gal) followed by THF (40 gal). The
solid was added to a solution of NaOH (6.02 kg, 150 6 mol)
in water (53 gal) and stirred for 60 min at 25 °C. The hazy
mixture was filtered and the pH of the filtrate adjusted to
7.0-7.2 with 3 N HCl. The resulting slurry was cooled to
10 °C and stirred for 1 h. The solids were filtered and washed
with water (53 gal) and THF (53 gal) and dried at 60 °C
under vacuum for 4 days, providing 32.2 kg (64.5%) of
zwitterion 1 as an off-white solid.
1
enantiomer. [R]D -52.3° (c ) 1.04, CHCl3). HMR (300
MHz, CDCl3) δ 8.64 (s, 1), 8.23 (s, 1), 8.17 (d, 1), 7.69 (d,
1), 7.14 (d, 2), 6.86 (d, 2), 5.48 (bs, 1), 4.86 (m, 1), 4.06 (t,
2), 3.50 (s, 2), 3.01 (t, 2), 2.90 (t, 1), 2.74 (m, 4), 2.20 (s,
3), 0.90 (s, 9), 0.1 (s, 3), -0.4 (s, 3). Mass spectrum: m/e:
500 (M+).
2-(4-{2-[2-(6-Acetylamino-pyridin-3-yl)-(2R)-hydroxy-
ethylamino]ethoxy}phenyl)-N-methyl-acetamide Dihydro-
chloride (8). A mixture of 12 (46.4 kg, 99.86 mol), amine
7 (41.6 kg, 199.7 mol), DMSO (8 gal), and N,N-diisopro-
pylethylamine (4.6 gal) was heated at 75 °C for 14 h, at
which time the reaction was judged complete by TLC
analysis (3:1, ethyl acetate/hexanes). After cooling to 35 °C,
water (39 gal) and ethyl acetate (54 gal) were added, and
the mixture stirred for 15 min at 25 °C. The agitation was
stopped and the mixture allowed to separate. The organic
layer was washed with water (2 × 11 gal) and concentrated
under vacuum to an oil. Toluene (20 gal) was added to the
oil, and the resulting solution was concentrated again under
vacuum to a stirrable oil. Tetrahydrofuran (86 gal) was added,
followed by a 1 M solution of tetrabutylammonium fluoride
in THF (34.3 gal, 130 mol), and the mixture stirred for 12
h at 25 °C. HPLC analysis indicated the reaction was
complete. To this solution was added ethanolic HCl (prepared
by adding acetyl chloride (5.6 gal, 299 mol) to absolute
ethanol (16 gal)), keeping the temperature between 5 and
10 °C during the addition. The resulting slurry was stirred
at 10 °C for 2 h, then filtered on a Rosenmund filter under
N2. The filter cake was washed successively with THF (15
gal), then diisopropyl ether (123 gal), and allowed to pull
dry for 60 min under N2. Acetonitrile (74 gal) was added to
the filter cake in the Rosenmund filter. The resulting slurry
was stirred at 25 °C for 12 h, filtered, and washed with
acetonitrile (78 gal) followed by isopropyl ether (61 gal).
The solid was dried under vacuum in the Rosenmund filter
at 45 °C for 12 h, affording 38.8 kg (84.7%) of the bis-HCl
salt of 8 as a nonhygroscopic pale yellow solid.
A slurry of the zwitterion 1 (32.2 kg, 97.16 mol) in USP-
grade water was stirred at 50 °C for 15 h, then cooled to 25
°C, and filtered. The filter cake was washed with USP-grade
water (17 gal) and speck-free THF (2 × 25 gal) and dried
under vacuum at 50 °C for 2-3 days, yielding 30.3 kg
(94.1%) of 1 as an ash-free white solid. The opposite
enantiomer was not detected by HPLC analysis (<0.05%).
1HMR (300 MHz, D2O + DCl) δ 7.93 (d, 1), 7.86 (s, 1),
7.28 (d, 2), 7.05 (d, 1), 7.00 (d, 2), 5.10 (dd, 1), 4.34 (t, 2),
3.69 (s, 2), 3.60 (t, 2), 3.40 (m, 2).
(4-{2-[2-(6-Amino-pyridin-3-yl)-(2R)-hydroxy-
ethylamino]ethoxy}phenyl)acetic Acid (1). The bis-HCl salt
of 8 (74.6 kg, 162.4 mol) was dissolved in water (83 gal),
and a solution of sodium hydroxide (32.5 kg, 812 mol) in
water (34 gal) was added over 20 min, keeping the reaction
temperature at 15-20 °C. Following the addition the thin
slurry was heated to 100 °C for 12 h, then cooled to
approximately 35 °C and sampled for reaction completion.
HPLC analysis indicted that the reaction was complete. Darco
Anal. Calcd for C17H21N3O4: C, 61.61; H, 6.40; N, 12.68.
Found: C, 61.49; H, 6.29; N, 12.60.
Received for review January 20, 2004.
OP049969O
(8) The pH adjustment took up to 12 h to stabilize at pH 7-7.2. Typically, the
pH was adjusted to approximately 8.5, upon which it would drift downward
into the desired range then back up again until eventually stabilizing in the
desired range.
(7) Chiral HPLC performed on Chiralpak AD column 4.6 cm × 25 cm, mobile
phase: 80:20 (hexane/isopropyl alcohol), 1.5 mL/min.
586
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development