Development of an enantioselective hydrogenation based synthesis of a glucokinase activator

Article abstract of DOI:10.1021/op300053a

This article describes the development and optimization of chemical reactions and subsequent multikilogram preparation of the glucokinase activator (R)-1 to fund clinical evaluation as a potential therapeutic for type II diabetes. The major process developments presented here are a Wittig olefination isomerization based synthesis of an E-acrylic acid, an optimized enantioselective hydrogenation of the E-acrylic acid, and a challenging final amide coupling.

Full text of DOI:10.1021/op300053a

Communication
pubs.acs.org/OPRD
Development of an Enantioselective Hydrogenation Based Synthesis
of a Glucokinase Activator
Nicholas A. Magnus,* Timothy M. Braden, Jonas Y. Buser, Amy C. DeBaillie, Perry C. Heath,
Christopher P. Ley, Jacob R. Remacle,‡ David L. Varie, and Thomas M. Wilson
Eli Lilly and Company, Chemical Product Research and Development Division, Indianapolis, Indiana 46285, United States
S
* Supporting Information
devised by Prosidion, the originator of (R)-1 (Scheme 1).2 The
ABSTRACT: This article describes the development and
optimization of chemical reactions and subsequent multi-
kilogram preparation of the glucokinase activator (R)-1 to
fund clinical evaluation as a potential therapeutic for type
II diabetes. The major process developments presented
here are a Wittig olefination isomerization based synthesis
of an E-acrylic acid, an optimized enantioselective
hydrogenation of the E-acrylic acid, and a challenging
final amide coupling.
Prosidion synthesis illustrated in Scheme 1 begins with a
Friedel−Crafts acylation which involves reacting cyclopropyl-
phenylsulfide3 (2) with ethyl 2-chloro-2-oxoacetate activated by
aluminum chloride (AlCl3) in CH2Cl2 (CH2Cl2 was necessary
for compatibility with the AlCl3 promoted chemistry) to afford
sulfide-α-ketoester (3) as a yellow oil with excellent
regiocontrol and good yield. For the following oxidation of
sulfide to sulfone, m-chloroperbenzoic acid (m-CPBA) was
found to be one of the most productive oxidants for producing
sulfone-α-ketoester 4. However, m-CPBA is relatively ex-
pensive, is potentially explosive,4 can be contaminated with
H2SO4, which promotes Baeyer−Villiger rearrangement to give
anhydride 5, and was shown to operate most effectively in
CH2Cl2 (toxicity of chlorinated solvents is undesirable) to
transform 3 into 4. Sulfone-α-ketoester 4 is transformed to the
E-acrylic acid (8) via Wittig olefination followed by double
bond isomerization and ester hydrolysis in poor overall yield.
For the Wittig olefination, phosphonium iodide (6)2b was
deprotonated with lithium hexamethyldisilazide in THF to
generate the corresponding ylide as a thin slurry and the
sulfone-α-ketoester 4 was added to it. This mode of addition
appeared problematic, as the initial low concentrations of
electrophile 4 bearing acidic protons in the basic ylide media
give rise to potential side reactions. E-Acrylic acid (8) was
converted to (R)-9 via enantioselective hydrogenation medi-
ated by the catalyst system formed from the precatalysts (R)-
(S)-MOD-Mandyphos and bis-(norbornadiene)-rhodium(I)-
tetrafluoroborate ([Rh(NBD)2]BF4).2b,5 The hydrogenation
required 750 psig of hydrogen and a substrate to catalyst
ratio (s/c) of 2000 to 1 to afford chiral acid (R)-9 with 86% ee.
The catalyst formed and the reaction ran best in MeOH, and
about 16% toluene by volume was added to aid the dissolution
of the E-acrylic acid (8). The hydrogenation afforded an 86% ee
of (R)-9 at a reaction temperature of 30 °C, which took about
18 h to reach completion (higher temperatures caused a decline
in ee). In addition, a charcoal adsorbent treatment was required
to remove residual rhodium, and a recrystallization to elevate
the ee of isolated (R)-9 to >99%. The final step is an amide
coupling with activated (R)-9 and poorly nucleophilic 2-
aminopyrazine 10. (R)-9 is activated via reaction with
stoichiometric Vilsmier reagent in CH2Cl2 and the resulting
mixture cooled to −45 °C before adding pyridine and 10 to the
reaction mixture to give (R)-1 (Note: CH2Cl2 was required for
INTRODUCTION
The glucokinase activator (GKA) (R)-1 was developed as a
potential therapeutic for the treatment of type 2 diabetes. Type
■
2 diabetes, a disease characterized by elevated blood glucose
concentrations, i.e., hyperglycemia, is becoming ever more
prevalent as a result of the recent dramatic rise in obesity
levels.1a The glucose-phosphorylating enzyme glucokinase
(GK) represents an attractive target for type 2 diabetes
therapies1b because it plays a critical role in whole-body glucose
control through its actions in multiple organs.1c−f In particular,
in the cells of the pancreas, GK acts as the glucose sensor that
determines the threshold for insulin secretion, while, in the
liver, this enzyme is rate-determining for glucose metabolism.
These GK activators have been shown to engender potent
antihyperglycemic actions in rodents, both by increasing
pancreatic insulin secretion and by augmenting hepatic glucose
metabolism.1g,h In effect, these dual pancreatic and hepatic
actions represent a “double whammy” on the hyperglycemia
associated with type 2 diabetes, and it may be that GK
activators could provide enhanced glycemic control by
combining, in a single molecule, the glucose-lowering effects
of insulin secretagogues, e.g., sulfonylureas, with hepatic
antidiabetic actions reminiscent of the biguanide metformin.
Due to the clinical trial requirements to evaluate GKA (R)-1,
we needed to develop a process capable of delivering ton
quantities of the API. Prior to our involvement with the project,
kilogram quantities of (R)-1 had been prepared via a synthesis
Received: March 2, 2012
Published: April 11, 2012
© 2012 American Chemical Society
830
dx.doi.org/10.1021/op300053a | Org. Process Res. Dev. 2012, 16, 830−835

Organic Process Research & Development
Communication
Scheme 1. Prosidion’s Synthesis of (R)-1
Scheme 2. Lilly’s Optimized Synthesis of (R)-1
solubility, and higher temperatures and bases stronger than
pyridine caused stereochemical erosion). When repeating this
reaction, we observed the imide 11 to form as a byproduct by
MS, as well as the trace byproduct 12 (see Supporting
Information for structural proof), and we concluded that
adding 10 to activated (R)-9 created a situation of an initial low
concentration of 10 which caused per-acylated byproducts to
form. Perhaps most problematic was the lack of a crystal form
and therefore purity control of (R)-1.
chemicals were changed and the reaction conditions were
optimized. A reliable and sustainable oxidation to convert
sulfide-α-ketoester 3 to sulfone-α-ketoester 4 was not identified
by Prosidion, so we decided to examine running the Wittig
olefination first followed by oxidation of sulfide to sulfone to
avoid the Baeyer−Villiger degradation issues. Wittig olefination
of 3 was conducted by generating the ylide of 6 as before, but
reversing the addition order to add the thin slurry of ylide to
the sulfide-α-ketoester 3. This afforded solution yields of E/Z-
sulfide-acrylates 13 of 80% (the opposite addition order gave
solutions yields of about 10% lower). We investigated the
isomerization of the mixture of E/Z-sulfide-acrylates 13 to the
E-isomer and determined that the sulfone functional group was
RESULTS AND DISCUSSION
■
To develop the Scheme 1 chemistry to prepare multi-kilogram
amounts of high purity (R)-1, the order of the steps and
831
dx.doi.org/10.1021/op300053a | Org. Process Res. Dev. 2012, 16, 830−835

Organic Process Research & Development
Communication
essential for a productive isomerization (see Supporting
Information for an illustration of an isomerization study).
The isomerization to the E-isomer was imperative to maximize
the yield of 8, since crystallization of acrylic acid 8 from MeOH
affords the E-isomer preferentially as a MeOH solvate, leaving
proportionately more Z-isomer than E-isomer in the liquor.
These observations dictated the Scheme 2 telescoped sequence
of Wittig olefination, sulfide to sulfone oxidation, olefin
isomerization, and ester hydrolysis. This telescoped process
from 3 to 8 without intermediate purifications was also
necessary, since triphenylphosphine oxide (Ph3PO), which was
generated in the first step Wittig chemistry, could not be
efficiently removed until production of acrylic acid 8, which
afforded Ph3PO removal via selective extraction. Therefore,
crude E/Z-sulfide-acrylates 13 contaminated with Ph3PO was
reacted with oxone in aqueous acetone to afford E/Z-7 in 84%
yield. Crude E/Z-7 contaminated with Ph3PO was reacted with
aqueous NaOH in MeOH to affect olefin isomerization to the
E-isomer and hydrolysis to E-acrylic acid (8), which was
worked up with a selective base extraction to remove Ph3PO
and crystallized from MeOH to give 8 as a MeOH solvate in
65% yield with >99% purity.
Chiral acid (R)-9 has about 20 times greater solubility in
MeOH than E-acrylic acid (8); therefore, it appeared practical
that the hydrogenation conversion could start as a suspension
and homogenize during the process. Indeed, this is the case,
and conversion of 8 into 9 transpires in just 3 h at 700 psig in a
neat MeOH system. The highly pure E-acrylic acid (8),
produced from the Scheme 2 synthesis, allowed the catalyst
load to be lowered to a s/c of 4000 to 1, obviating the need for
a workup and adsorbent treatment to remove rhodium. The
chiral acid (R)-9 was crystallized from a mixture of 2-MeTHF
and heptane to afford an 87% yield with >99% ee.
In our attempts to find a solvent system for the hydro-
genation that would improve the solubility of E-acrylic acid (8),
we found that the triethylammonium carboxylate salt of 8
showed excellent solubility in MeOH (>1 g/mL). In addition,
TEA gave the benefit of dramatically increasing reduction rates,
to the point that the hydrogen pressure could be reduced to 70
psig, with complete reductions transpiring in less than 4 h (this
procedure required a workup involving an acid wash to remove
the TEA).6,7 Interestingly, reduction rates began to decrease
when greater than 1 equiv of TEA was used. The (R)-(S)-
MOD-Mandyphos ligand contains basic amine functionality;
hence, we speculate that the triethylamine may either be
assisting with dissociation of 9 from the catalyst complex or
relieving ligand protonation, leading to improved reduction
efficiency. Triethylamine did not impact the enantioselectivity
of the reduction, and (R)-9 was produced with 89% ee and was
isolated via the aforementioned crystallization in 86% yield with
>99% ee.
For the final step, amide coupling, the order of addition was
reversed from the Prosidion procedure so that the activated
chiral acid (R)-9 was added to the 2-aminopyrazine (10). This
addition order did mitigate per-acylated byproducts, and as an
added bonus, the stereocenter became stable toward epimeriza-
tion under ambient conditions. This equated to a 65 °C
increase in reaction temperature (from −45 to 20 °C) and
made it possible to switch the solvent system from CH2Cl2 to
THF due to increased solubility. The final procedure was to
dissolve (R)-9 in THF at 20 °C, and add catalytic DMF,
followed by slow addition of oxylyl chloride to produce the
corresponding acid chloride. The acid chloride was then added
to a second reactor containing 2-aminopyrazine (10) in THF
with 10 equiv of pyridine present to minimize formation and
precipitation of the hydrochloride salt of 10. This procedure
resulted in solution yields of 89−94% of (R)-1 with the
stereocenter maintained with >99% ee. After reaction
completion, pyridine hydrochloride was filtered off, EtOAc
was added to improve phase separations, and residual 10,
pyridine, and 4−5% of unreacted (R)-9 were washed away with
acid and base, respectively. The resulting solvents in the
solution of crude (R)-1 in THF/EtOAc were replaced with
MeOH via azeotropic distillation and seeding employed to
crystallize a form of (R)-1 that was discovered at Lilly in 73%
yield.8 Attempts to improve the yield by reducing the solubility
of (R)-1 in the crystallization system led to production of crude
solids with darker color and poor physical properties.
CONCLUSION
■
To avoid Baeyer−Villiger degradation issues, permit use of
more benign reagents and solvents, and improve control, yield,
and purity, the steps for the preparation of E-acrylic acid (8)
were developed and the sequence changed. The high quality E-
acrylic acid (8) produced by the new process led to improved
productivity in the subsequent high pressure enantioselective
hydrogenation to produce chiral acid (R)-9. In addition, low
pressure conditions for the hydrogenation involving in situ
formation of an ammonium carboxylate of E-acrylic acid (8)
were identified. Lastly, development of the final amide coupling
and crystallization resulted in a six step synthesis starting from
cyclopropylphenyl sulfide 2 to produce the GKA, (R)-1 in 23%
overall yield.
EXPERIMENTAL SECTION
General. H and 13C NMR spectra were obtained on a
■
1
Varian spectrometer at 400 and 101 MHz, respectively.
Processes for the preparation of cyclopropylphenyl sulfide (2)
and triphenyl((tetrahydro-2H-pyran-4-yl)methyl)phosphonium
iodide (6) have not been investigated or developed at Lilly.
Cyclopropylphenyl sulfide (2) and triphenyl((tetrahydro-2H-
pyran-4-yl)methyl)phosphonium iodide (6) are known com-
pounds, and the respective preparations have been docu-
mented.2,3 Spectroscopic data for 6: H NMR (DMSO-d6, 400
1
MHz) 7.83−7.89 (m, 9H), 7.72−7.77 (m, 6H), 3.62−3.67 (m,
4H), 3.12 (dt, 2H, J = 2.8, 11.6 Hz), 1.90−1.99 (m, 1H), 1.30−
1.42 (m, 4H). 13C NMR (DMSO-d6, 101 MHz) 135.39,
135.36, 134.1, 134.0, 130.7, 130.6, 119.8, 119.0, 66.7, 34.0, 33.9,
30.2, 30.1, 26.9, 26.4. IR (film) 2927, 2834, 1432 cm−1. HRMS
(ESI+) caldc for C24H26IOP 488.07659, found 488.07643.
Analytical methods: GC: Column: DB-5 MS; 30 m × 0.25
mm, 0.25 μm; Carrier gas: helium; Flow rate: 2.3 mL/min;
Temp: 50 °C, hold for 2.0 min, then ramp at 20 °C/min to 260
°C; Run time: 22.5 min. HPLC for enantiomeric excess
determination of chiral acid (R)-9: Column: ChiralPak AD-RH,
(0.46 cm × 15 cm, 5 μm), Flow rate: Isocratic flow, 0.6 mL/
min, 30 °C, UV detection: 230 nm. Eluent: 68% H2O/32%
CH3CN/0.1% TFA. Run time: 15 min, (R)-9 rt = 7.8, (S)-9 rt
= 10.9. HPLC for enantiomeric excess determination of chiral
amide (R)-1: Column: Diacel OJ-RH (0.46 cm × 15 cm, 5
μm), Flow rate: Isocratic flow, 0.5 mL/min., 23 °C, UV
detection: 235 nm. Eluent: 65% H2O/35% CH3CN. Run time:
30 min, (R)-1 rt = 9.1, (S)-1 rt = 12.7. HPLC for reaction
monitoring: Colunm: Agilent Ecilipse XDB-C8 (0.46 cm × 15
cm, 3.5 μm). Flow rate: 2 mL/min, A = 0.1% H3PO4, B =
832
dx.doi.org/10.1021/op300053a | Org. Process Res. Dev. 2012, 16, 830−835

Organic Process Research & Development
Communication
CH3CN, Gradient: 95% A to 5% A over 10 min; change to 95%
A over 1 min and hold for 4 min, 30 °C, UV detection: 220 nm.
Ethyl 2-(4-(Cyclopropylthio)phenyl)-2-oxoacetate (3).
A reactor was charged with CH2Cl2 (702.0 kg) and cooled to
0−5 °C, and AlCl3 (109.5 kg, 820.2 mol) was added to the
mixture in five portions while maintaining the temperature at
0−5 °C. After the addition was complete, the mixture was
stirred for 0.25 h, and ethyl 2-chloro-2-oxoacetate (90.0 kg,
659.3 mol) was added to the mixture at a rate of 18−22 kg/h at
0−5 °C. After the addition, the temperature was maintained at
0−5 °C for 1.5−2.0 h. Maintaining the temperature at 0−5 °C,
cyclopropylphenyl sulfide 2 (90.0 kg, 599.2 mol) was added to
the mixture at a rate of 18−22 kg/h. After 2.0 h, a sample of the
reaction mixture was analyzed by GC, which indicated that the
cyclopropylphenyl sulfide 2 was <3.0%. The reaction mixture
was transferred to a second reactor at a rate of 200−250 kg/h
in five portions into crushed ice (900.0 kg) and water (270.0
kg) while maintaining the temperature at 0−10 °C. After
quenching the reaction, the mixture was held for 0.5 h at 5−10
°C. The aqueous phase was separated and extracted with
CH2Cl2 (282.0 kg). The organic phases were combined and
washed with a 6.3% aqueous solution of NaHCO3 (302.4 kg),
followed by saturated brine (162.4 kg) until the KF
measurement of the organic layer was <0.3%. The organic
phase was concentrated under atmospheric pressure at 39−45
°C until no further distillate could be collected, and then the
mixture was concentrated under vacuum (140 mmHg) below
35 °C until the CH2Cl2 content was <1.0%. The product
sulfide-α-ketoester 3 was a yellow liquid which had a weight of
139.0 kg with a 92.3% potency (yield 85.5%). 1H NMR
(CDCl3, 400 MHz) 7.90 (d, 2H, J = 8.8 Hz), 7.43 (d, 2H, J =
8.8 Hz), 4.42 (q, 2H, J = 7.2 Hz), 2.18 (tt, 1H, J = 4.4, 7.6 Hz),
1.40 (t, 3H, J = 7.2 Hz), 1.13−1.18 (m, 2H), 0.69−0.73 (m,
2H). 13C NMR (CDCl3, 101 MHz) 185.5, 163.8, 149.7, 130.0,
128.5, 125.1, 62.1, 14.0, 10.8, 8.5. IR (film) 1729, 1670, 1584,
1283, 1205, 1175 cm−1. HRMS (ESI+) calcd for C13H14O3S
250.06637, found 250.06613.
(E/Z)-Ethyl 2-(4-(Cyclopropylthio)phenyl)-3-(tetrahy-
dro-2H-pyran-4-yl)acrylate (13). A reactor was charged
with THF (355.0 kg) followed by hexamethyldisilazane (135.4
kg, 838.9 mol). The mixture was cooled to −10 to 0 °C, and n-
butyllithium (246.1 kg, 881.5 mol) was added at a rate of 25−
30 kg/h while maintaining the temperature between −10 and 0
°C. After the addition was complete, the temperature was
maintained at −10 to 0 °C for 2.0−3.0 h. To a second reactor
was charged THF (533.0 kg) followed by phosphonium iodide
(6; 389.5 kg, 798.2 mol) in four portions. The lithium
hexamethyldisilazide/THF mixture was added to the phospho-
nium iodide (6)/THF mixture at a rate of 80−110 kg/h at 0−
10 °C, and after the addition, the temperature was maintained
at 0−10 °C for 0.5 h. The temperature was raised to 15−20 °C,
and after 2.5 h the mixture was cooled to 0−5 °C, and in a third
reactor was charged THF (355.0 kg) followed by the addition
of sulfide-α-ketoester 3 (200.6 kg, 801.4 mol). The mixture was
cooled to 0−10 °C, and the ylide mixture was added to the
sulfide-α-ketoester 3 mixture within 8.0−10.0 h while
maintaining the temperature at 0−10 °C. After the addition
was complete, the reaction mixture was maintained at 15−20
°C, and after 2.0 h a sample of the mixture was analyzed by
HPLC, which indicated that the sulfide-α-ketoester 3 was
<2.0%. The resulting mixture was transferred to a reactor that
contained an aqueous 1 M solution of citric acid (1006.0 kg)
(final pH = 2.0−3.0) while maintaining the temperature at 0−
15 °C. The mixture was filtered, and the filter cake (Ph3PO)
was rinsed with THF (10.0 kg). The aqueous phase was
separated, and the organic phase was concentrated under
atmospheric pressure at 65−70 °C until no further distillate
could be removed. The concentrated organic phase was cooled
to 20−25 °C, and the aqueous phase was extracted with MTBE
(295.0 kg) and separated. The separated organic phase was
added to the concentrated organic phase, and the resulting
mixture was concentrated at atmospheric pressure at 56−60 °C
until no further distillate could be removed. The organic phase
was further concentrated under vacuum (140 mmHg) at 40−45
°C until no further distillate could be removed, and the organic
phase was left under vacuum for a further 6.0−8.0 h until the
THF and MTBE contents were <10.0%. The organic phase was
cooled to 20−25 °C, and acetone (100.0 kg) was added
followed by stirring for 0.5 h. The acetone solution weighed
213.2 kg and contained Ph3PO and a ∼1:1 mixture of E:Z-
sulfide acrylic ester product 13, with 41.5% potency (yield
80.0%), which was taken directly into the next step.
Ethyl 2-(4-(Cyclopropylsulfonyl)phenyl)-3-(tetrahy-
dro-2H-pyran-4-yl)acrylate (7). A reactor was charged with
a solution of sulfide 13 in acetone (213.2 kg; 41.5 wt %; 641.2
mol), from the prior step, followed by the addition of water
(213.2 kg) at 5−15 °C. Oxone (K2SO4, KHSO4, 2KHSO5)
(650.2 kg, 2116.1 mol) was added to the mixture in 23 portions
over 11.5 h at 5−15 °C. The mixture was analyzed by HPLC,
the peak area of 13 (E+Z) was <0.1%, and intermediate content
(E+Z-sulfoxide) was <0.3%. While maintaining the reaction
temperature between 0 and 15 °C, saturated aqueous NaHSO3
solution (336.8 kg) was added to the mixture at a rate of 50.0−
60.0 kg/h until potassium iodide-starch paper did not turn blue.
The mixture was transferred to a centrifuge for filtration, and
the filter cake (Ph3PO) was washed with acetone (2 × 341.0
kg) until the product content in the filter cake was <0.5%. The
filtrate was transferred to a reactor, and 2 M NaOH solution
(1302.0 kg) was added to the mixture at a rate of 140−160 kg/
h until the pH was 6−7. The mixture was concentrated under
vacuum (140 mmHg) at 40−47 °C until no more acetone
distilled out. The mixture was cooled to 25−30 °C and
extracted with MTBE (471.2 kg), and the aqueous phase was
separated. The aqueous phase was extracted with MTBE (471.2
kg), and the product content in the aqueous phase was <0.5%.
The organic phases were combined and washed with water (2
× 213.2 kg). The organic phase was concentrated under
atmospheric pressure at 55−62 °C until no further distillate
could be removed. The organic phase was concentrated under
vacuum (140 mmHg) at 30−40 °C until no further distillate
could be removed, and the residue continued to concentrate
under vacuum for 9.0 h until the MTBE and acetone content
was <10.0%. The residue was cooled to 20−25 °C, and MeOH
(336.8 kg) was added to the mixture (E/Z-7 and Ph3PO) for
the following step. The MeOH solution of E/Z-7 had a weight
of 197.0 kg, had a potency of 26.8% (yield 84.3%), appeared as
a yellow liquid, and was taken directly into the next step.
(E)-2-(4-(Cyclopropylsulfonyl)phenyl)-3-(tetrahydro-
2H-pyran-4-yl)acrylic Acid (8). The solution of E/Z-7 (197.0
kg, 541.2 mol) in MeOH (624.0 kg), from the prior step, was
charged to a reactor. To the solution was added a 7.4% aqueous
solution of NaOH (779.5 kg) while maintaining the temper-
ature between 10 and 30 °C. The mixture was heated to 65−70
°C for 2.0 h, and a sample of the reaction mixture was analyzed
by HPLC, which indicated that 7 (E+Z) was <0.3%. The
mixture was cooled to 40−45 °C and concentrated under
833
dx.doi.org/10.1021/op300053a | Org. Process Res. Dev. 2012, 16, 830−835

Organic Process Research & Development
Communication
vacuum (140 mmHg) at 43−50 °C until no further distillate
was observed (content of methanol in the aqueous phase was
<3.0%). The resulting mixture was cooled to 10−15 °C, and
water (1383.0 kg) was added. After the addition was complete,
the temperature was maintained at 10−15 °C for 2.0−3.0 h, the
reaction mixture was filtered by centrifuge, and the filter cake
(Ph3PO) was washed with a 3.7% aqueous solution of NaOH
(296.4 kg), and water (296.0 kg), respectively. All of the
filtrates were combined and extracted with MTBE (449.0 kg +
298.0 kg). The organic phases were combined and washed with
water (198.0 kg). All of the aqueous phases were combined,
and MTBE (449.0 kg) was added while maintaining the
temperature at 13−18 °C. The pH of the mixture was adjusted
to 1−2 with 1 M sulfuric acid (1183.0 kg) at 13−18 °C. After
the addition was complete, the temperature was maintained at
13−18 °C for 5.0−6.0 h. The mixture was filtered by centrifuge,
and the filter cake was washed with water (2 × 84.0 kg). The
filter cake was dried at 35−40 °C until the KF measurement
was <3.0%, and crude E-acrylic acid (8; 132.0 kg) was obtained.
The aqueous phase of the filtrate was separated and extracted
with MTBE (2 × 449.0 kg). The organic phases were
combined, and active carbon (9.9 kg) was added, which was
rinsed with MTBE (25.0 kg) in advance. The mixture was
heated to 40−50 °C and the temperature maintained at 40−50
°C for 1.0 h. The mixture was filtered at 20−40 °C and
concentrated under vacuum (140 mmHg) at 35−45 °C until
there was 150−200 L remaining. MTBE (126.0 kg) was added
to the mixture, stirred for 1.0 h, and cooled to 0−5 °C for 3.0−
4.0 h. The mixture was filtered by centrifuge, and product was
reclaimed from the mother liquor again to give crude 8 (15.2
kg). The crude 8 (132.0 kg +15.2 kg) was recrystallized from
MeOH (927.4 kg). The mixture was heated to 63−68 °C and
stirred until the solids were completely dissolved. The mixture
was cooled to 10−15 °C within 6.0 h and the temperature
maintained at 10−15 °C for 1.0 h. The mixture was cooled to
0−5 °C within 2.0 h and the temperature was maintained at 0−
5 °C until the wt % of 8 was <2.0% in the filtrate. The mixture
was filtered by centrifuge, and the filter cake was washed with
cold (0−5 °C) MeOH (58.9 kg). The cake was dried at 35−40
°C until the KF measurement was <0.5% and the LOD
measurement was <3.0%. E-Acrylic acid (8) was isolated as a
MeOH solvate and appeared as an off-white solid with a weight
of 132.9 kg, a potency of 88.9%, and a purity of 99.9% (yield
64.9%). Mp 152.5−153.6 °C. 1H NMR (DMSO-d6, 400 MHz)
12.72 (s, 1H), 7.87 (d, 2H, J = 8.4 Hz), 7.43 (d, 2H, J = 8.4
Hz), 6.81 (d, 1H, J = 10.3 Hz), 3.70−3.80 (m, 2H), 3.12−3.18
(m, 2H), 2.88−2.91 (m, 1H), 2.12−2.26 (m, 1H), 1.38−1.55
(m, 4H), 1.04−1.18 (m, 4H). 13C NMR (DMSO-d6, 101 MHz)
167.7, 148.4, 141.4, 139.9, 132.3, 131.0, 127.4, 66.3, 35.7, 32.4,
31.4, 5.9. IR (neat) 3497, 2940, 2910, 2612, 1689, 1629, 1414,
1391. HRMS (ESI+) calcd for C17H21O5S 337.1104, found
337.1099.
process was repeated 4 times). A transfer line to add the
catalyst solution to the autoclave was purged with N2, and then
the catalyst mixture was transferred to the autoclave by utilizing
pressure. The autoclave was pressurized to 700 psig with H2,
and the reaction mixture was heated to 30 °C for 3.0 h (the
reaction starts as a slurry and becomes homogeneous). The
autoclave was cooled to 20 °C, vented to 5−20 psig, and
sampled and assayed for reaction completion by HPLC, which
indicated 0.11% of 8 remaining. The reaction mixture was
transferred to a reactor with MeOH (5 L) used to rinse the
autoclave, which was added to the reactor. The resulting
mixture was distilled at 23 °C under vacuum to a final volume
of approximately 24 L. 2-MeTHF (84 L) was added to the
mixture and distillation continued to a final volume of 60 L.
Additional 2-MeTHF (24 L) was added to the mixture during
distillation at the same rate as distillate removal while
maintaining the reactor volume at 60 L. The mixture was
heated to 50 °C, and heptane (24 L) was added over 0.5 h.
Seeds of (R)-9 (10 g) were suspended in heptane (200 mL)
and added to the reaction mixture, followed by a heptane (36
L) addition over 1.0 h. The resulting slurry was cooled to 20 °C
over 2.0 h, stirred for 3.0 h, and filtered. The filter cake of (R)-9
was washed with a mixture of 2-MeTHF (10 L) and heptane
(23 L), and the cake was dried with a stream of N2 for 1.0 h and
further dried via roto-vap under vacuum at 50−60 °C. The
chiral acid (R)-9 was isolated as an off white solid with a weight
of 10.43 kg, potency of 99.8%, purity of 99.9%, and chiral purity
of 99.72% ee with 29 ppm of Rh (yield 87.5%). Mp 153.9−
1
154.4. H NMR (DMSO-d6, 400 MHz) 0.99−1.05 (m, 2H),
1.06−1.19 (m, 4H), 1.26−1.33 (m, 1H), 1.50−1.60 (m, 2H),
1.63 (ddd, 1H, J = 7.2, 7.2, 14.0 Hz), 1.92 (ddd, 1H, J = 7.6, 7.6,
13.6 Hz), 2.82 (dddd app tt, 1H, J = 4.8, 8.0 Hz), 3.11−3.18
(m, 2H), 3.73−3.80 (m, 3H), 7.57 (d, 2H, J = 8.4 Hz), 7.83 (d,
2H, J = 8.4 Hz), 12.54−12.60 (br s, 1H). 13C NMR (DMSO-d6,
101 MHz) 5.8, 32.4, 32.6, 32.8, 33.0, 48.0, 67.3, 127.9, 129.4,
139.6, 146.0, 174.6. IR (film) 3000−2600 (br s), 2916, 1718,
1316, 1294, 1272, 1223 cm−1. HRMS (ESI+) calcd for
24
C17H22O5S 338.1187, found 338.1183. [α]D −53.8 (MeOH,
c = 2.8).
Preparation of (R)-2-(4-(Cyclopropylsulfonyl)phenyl)-
3-(tetrahydro-2H-pyran-4-yl)propanoic Acid ((R)-9) in
the Presence of Triethylamine. To prepare the catalyst,
degassed MeOH (3.8 mL), under N2, was combined with (R)-
(S)-MOD-Mandyphos (SL-M004-1) (11.1 mg, 0.0105 mmol)
and [Rh(NBD)2]BF4 (3.7 mg, 0.01 mmol), and the resulting
mixture was stirred under ambient conditions (∼23 °C) for
14.0 h. In a separate reactor, E-acrylic acid (8; 13.35 g, 39.65
mmol) was combined with triethylamine (4.01 g, 39.65 mmol)
and MeOH (62.8 mL) to provide a colorless solution. The
solution was purged of O2 by pressure cycling with N2 to 50
psig (repeated 3 times). While maintaining an inert environ-
ment, the catalyst solution was transferred to the reactor and
the system was pressurized to 70 psig with H2 with a reactor
temperature of 20 °C. After 4.0 h, the reaction was sampled and
assayed for completion by HPLC, which indicated <1% of 8
remaining. The system was purged of H2, and then a portion of
the MeOH was removed by distillation under reduced pressure
(300 Torr, 42 °C) to give approximately 40 mL of reactor
volume. Water (18 mL) was added, and the remaining MeOH
was removed by distillation under reduced pressure (250 Torr,
55 °C), resulting in an approximately 25−28 mL reactor
volume. The condensate was diluted with 2-MeTHF (100 mL)
and then extracted with 1 M H2SO4 (20 mL). The organic
(R)-2-(4-(Cyclopropylsulfonyl)phenyl)-3-(tetrahydro-
2H-pyran-4-yl)propanoic Acid ((R)-9). To prepare the
catalyst, degassed MeOH (900 mL), under N2, was combined
with (R)-(S)-MOD-Mandyphos (SL-M004-1) (9.68 g, 9.17
mmol) and [Rh(NBD)2]BF4 (3.25 g, 8.69 mmol), and the
resulting mixture was stirred under ambient conditions (∼23
°C) for 14.0 h. In a separate reactor, E-acrylic acid (8; MeOH
solvate) (13.3 kg with 88.94% potency = 11.83 kg, 35.17 mol)
was suspended in MeOH (52 L) and charged to an autoclave
using vacuum for the transfer. The autoclave was pressurized
with N2 to 700 psig for >1 min and vented to 1−20 psig (this
834
dx.doi.org/10.1021/op300053a | Org. Process Res. Dev. 2012, 16, 830−835

Organic Process Research & Development
Communication
phase was concentrated (at atmospheric pressure, 71 °C) to a
reactor volume of 42 mL. To the hot solution was charged 20
mL of 2-MeTHF followed by 25 mL of heptane. After cooling,
the product was collected by filtration and then dried under
vacuum to give an 86% yield (11.55 g, 99.8% ee) of (R)-9
identical to that described above.
(dd, 1H, J = 1.2, 2.4 Hz), 9.30 (d, 1H, J = 1.2 Hz), 11.1 (s, 1H).
13C NMR (DMSO-d6, 101 MHz) 5.8, 32.4, 32.9, 33.0, 48.4,
67.2, 128.0, 129.3, 136.6, 139.7, 140.5, 143.1, 146.1, 149.0,
172.3. IR (film) 3239, 2841, 1670, 1532, 1405 cm−1. HRMS
(ESI+) calcd for C21H25N3O4S 415.15658, found 415.1562.
24
[α]D −49.5 (MeOH, c = 3.2).
(R)-2-(4-(Cyclopropylsulfonyl)phenyl)-N-(pyrazin-2-
yl)-3-(tetrahydro-2H-pyran-4-yl)propanamide ((R)-1). At
20−25 °C, a reactor was charged with THF (100 L) followed
by chiral acid (R)-9 (23.10 kg, 55.59 mol) which was rinsed
into the reactor with THF (5 L). DMF (0.275 kg, 3.76 mol)
was charged to the reactor followed by the addition of oxalyl
chloride (9.125 kg, 71.89 mol) over 0.25 h, maintaining the
temperature below 30 °C. The resulting mixture was held for
1.0 h, and a sample was quenched into MeOH for HPLC
analysis, which indicated >99% conversion to the acid chloride
via analysis of the methyl ester derivative. THF (116 L) was
charged to a second reactor followed by 2-aminopyrazine (10;
7.15 kg, 75.23 mol), which was rinsed into the reactor with
THF (6 L). Pyridine (54 kg, 682.7 mol) was charged to the
second reactor, and the temperature was maintained at 20−25
°C. The acid chloride mixture was transferred to the mixture of
10 in THF over 0.75 h while maintaining the temperature
below 30 °C. THF (24 L) was used to rinse the acid chloride
reactor into the second reactor. The resulting mixture was held
for 1.0 h, and a sample was quenched into MeOH for HPLC
analysis, which indicated <1% of the methyl ester derivative and
production of (R)-1. Solids (pyridinium hydrochloride) were
removed via filtration, and the resulting waste cake was washed
with EtOAc (230 L), which was combined with the reaction
mixture. Aqueous HCl (81.8 kg of 37% HCl in 61 L water) was
added to the reaction mixture over 10 min while maintaining
the temperature below 30 °C. At 20−25 °C, the aqueous layer
was separated and aqueous NaHCO3 (4.88 kg in 64.2 kg water)
was added to the reaction mixture over 8 min followed by
removal of the aqueous layer. The resulting mixture was heated
to 30−45 °C under vacuum (190−210 mmHg) to remove
distillate to a final volume of 90−95 L. EtOAc (230 L) was
added to the mixture followed by distillation at 30−45 °C
under vacuum (190−220 mmHg) to a final volume of 90−95 L
(this operation was repeated two more times to give 0.02%
water by KF and <1.0% THF by GC). MeOH (230 L) was
added to the reactor followed by distillation at 30−45 °C under
vacuum (170−220 mmHg) to a final volume of 115 L (this
operation was repeated two more times to a final volume of 95
L to give 0.14% water by KF and <1.0% EtOAc by GC). The
mixture was maintained at 38−42 °C while seed crystals of (R)-
1 (0.23 kg, 0.55 mol) were added. The resulting mixture was
held at 40 °C for 2.0 h, cooled to −10 °C over 8.0 h, and held
at −10 °C for 1.0 h. The solids were filtered on a filter dryer
and washed with cold (−10 °C) MeOH (92 L). The filter dryer
was maintained at 25 °C while N2 was blown through the
product cake for 2.0 h; then a switch to pulling vacuum through
the bottom of the cake was made until constant weight was
achieved. The chiral amide (R)-1 was isolated as an off white
solid with a weight of 21.06 kg, potency of 100.0%, and chiral
ASSOCIATED CONTENT
* Supporting Information
Structure proof for compound 12 and HPLC trace of 8's
transformation in 2 M methanolic NaOH. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*Phone, 317-651-1470; fax, 317-276-4507; e-mail, magnus_
nicholas@lilly.com.
■
Present Address
‡Dow Corning Corporation, 3901 S. Saginaw Rd., Midland, MI
48640, U.S.A.
Notes
The authors declare no competing financial interest.
REFERENCES
■
(1) (a) Moller, D. E. Nature 2001, 414, 821−827. (b) Sarabu, R.;
Tilley, J. Annu. Rep. Med. Chem. 2005, 40, 167−181. (c) Sarabu, R.;
Grimsby, J. Curr. Opin. Drug Discovery Dev. 2005, 8, 631−637.
(d) Kietzmann, T.; Ganjam, G. K. Expert Opin. Ther. Pat. 2005, 15,
705−713. (e) Guertin, K. R.; Grimsby, J. Curr. Med. Chem. 2006, 13,
1839−1843. (f) Johnson, T. O.; Humphries, P. S. Annu. Rep. Med.
Chem. 2006, 41, 141−154. (g) Grimsby, J.; Sarabu, R.; Corbett, W. L.;
Haynes, N.-E.; Bizzarro, F. T.; Coffey, J. W.; Guertin, K. R.; Hilliard,
D. W.; Kester, R. F.; Mahaney, P. E.; Marcus, L.; Qi, L.; Spence, C. L.;
Tengi, J.; Magnuson, M. A.; Chu, C. A.; Dvorozniak, M. T.;
Matschinsky, F. M.; Grippo, J. F. Science 2003, 301, 370−373.
(h) Coope, G. J.; Atkinson, A. M.; Allott, C.; McKerrecher, D.;
Johnstone, C.; Pike, K. G.; Holme, P. C.; Vertigan, H.; Gill, D.;
Coghlan, M. P.; Leighton, B. Br. J. Pharmacol. 2006, 149, 328−335.
(2) (a) Fyfe, M. C. T.; Gardner, L. S.; Nawano, M.; Procter, M. J.;
Rasamison, C. M.; Schofield, K. L.; Shah, V. K.; Yasuda, K. (OSI
Pharmaceuticals, Inc., USA; Prosidion Ltd, GB). PCT Int. Appl. 2004,
WO 2004072031 A2 20040826. (b) Briner, P. H.; Fyfe, M. C. T.;
Madeley, J. P.; Murray, P. J.; Procter, M. J. (Prosidion Ltd, GB);
Spindler, F. (Solvias, AG). PCT Int. Appl. 2006, WO 2006016178 A1
20060216.
(3) (a) Truce, W. E.; Hollister, K. R.; Lindy, L. B.; Parr, J. E. J. Org.
Chem. 1968, 33, 43. (b) Masson, E.; Leroux, F. Helv. Chim. Acta 2005,
88, 1375. (c) Angelauda, R.; Landais, Y. Tetrahedron 2000, 56, 2025.
(4) (a) Brougham, P.; Cooper, M. S.; Cummerson, D. A.; Heany, H.;
Thompson, N. Synthesis 1987, 1015. (b) Hazards in the Chemical
Laboratory; Luxon, S. G., Ed.; Royal Society of Chemistry: Cambridge,
1992.
(5) Spindler, F.; Malan, C.; Lotz, M.; Kesselgruber, M.; Pittelkow, U.;
Rivas-Nass, A.; Briel, O.; Blaser, H. Tetrahedron: Asymmetry 2004, 15,
2299.
(6) While we demonstrated at gram scale that the triethylammonium
salt of 8 can be converted to (R)-9 under low pressure conditions, this
chemistry was not developed for the pilot plant, since the project lost
funding.
(7) Yamamoto, K.; Ikeda, K.; Yin, L. K. J. Organomet. Chem. 1989,
370, 319. Sun, X.; Zhou, L.; Wang, C. J.; Zhang, X. Angew. Chem., Int.
Ed. 2007, 46, 2623. Fox, M. E.; Jackson, M.; Lennon, I. C.; Klosin, J.;
Abboud, K. A. J. Org. Chem. 2008, 73, 775. Li, S.; Zhu, S. F.; Zhang, C.
M.; Song, S.; Zhou, Q. L. J. Am. Chem. Soc. 2008, 130, 8584.
(8) Dunlap, J. T.; Stephenson, G. A. (Eli Lilly and Company). U.S.
Pat. Appl. Publ. US 20090181981 A1 20090716, 2009.
1
purity of >99% ee (yield 73.6%). Mp 156.8−157.9 °C. H
NMR (DMSO-d6, 400 MHz) 0.96−1.01 (m, 2H), 1.03−1.08
(m, 2H), 1.11−1.24 (m, 2H), 1.31−1.38 (m, 1H), 1.51−1.55
(m, 1H), 1.60−1.67 (m, 2H), 2.03−2.11 (m, 1H), 2.79 (tt, 1H,
J = 4.8, 8.0 Hz), 3.13−3.20 (m, 2H), 3.76 (ddd, 2H, J = 2.0, 2.0,
11.6 Hz), 4.21 (dd, 1H, J = 6.0, 8.4 Hz), 7.65 (d, 2H, J = 8.8
Hz), 7.84 (d, 2H, J = 8.4 Hz), 8.32 (d, 1H, J = 2.4 Hz), 8.35
835
dx.doi.org/10.1021/op300053a | Org. Process Res. Dev. 2012, 16, 830−835