Preparation of a HMG-CoA reductase inhibitor via an optimized imidazole-forming condensation reaction

Article abstract of DOI:10.1021/op800092e

Development work toward an enabling synthesis of preparative scale batches of an imidazole-based HMG-CoA reductase inhibitor is described. The desired target was synthesized in 16% yield over 7 steps, highlighted by an imidazole-forming condensation reaction in which the yield was improved from 20% to >70% via modification of the solvent, acid, and amine equivalents. The step 2 acylation was improved, and a problematic benzyl ester in step 4 was converted into the corresponding benzyl amide to decrease trans-amidation during the step 5 imidazole formation. A highly effective salt formation and crystallization protocol was also developed.

Full text of DOI:10.1021/op800092e

Organic Process Research & Development 2008, 12, 1183–1187
Preparation of a HMG-CoA Reductase Inhibitor via an Optimized Imidazole-Forming
Condensation Reaction
Daniel M. Bowles,* Gary L. Bolton, David C. Boyles, Timothy T. Curran, Richard H. Hutchings, Scott D. Larsen,
Jonathan M. Miller, William K. C. Park, Kurtis G. Ritsema, and David C. Schineman
Pfizer Global Research and DeVelopment, Michigan Laboratories, 2800 Plymouth Road, Ann Arbor Michigan 48105, U.S.A.
Markus Tamm
Carbogen AMCIS AG, Neulandweg 5, CH-5502 Hunzenschwil, Switzerland
Abstract:
Development work toward an enabling synthesis of preparative
scale batches of an imidazole-based HMG-CoA reductase inhibitor
is described. The desired target was synthesized in 16% yield over
7 steps, highlighted by an imidazole-forming condensation reaction
in which the yield was improved from 20% to >70% via
modification of the solvent, acid, and amine equivalents. The step
inhibitor.
2 acylation was improved, and a problematic benzyl ester in step
Figure 1. Imidazole 1, a potential HMG-CoA reductase
Scheme 1. Two approaches to imidazole 3a
4 was converted into the corresponding benzyl amide to decrease
trans-amidation during the step 5 imidazole formation. A highly
effective salt formation and crystallization protocol was also
developed.
Introduction
a Reagents and conditions: (a) NCCO2Bn, Ac2O, CF3C6H5, reflux, 24 h, 15%;
(b) 4, xylene, AcOH, TsOH (cat.), reflux, 16 h, 20%.
Statins, 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-
CoA) reductase inhibitors,1 are widely prescribed in the fight
against coronary heart disease2 for their ability to moderate
hypercholesterolemia.3 Unfortunately, many marketed statins,
albeit effective in lowering levels of LDL-C (so-called “bad
cholesterol”),4 have been linked to a variety of musculoskeletal
problems, such as cramping, myalgia,5 and rhabdomyolysis.6
Imidazole 1 (Figure 1) is a HMG-CoA reductase inhibitor with
potential for improved efficacy and tolerability7 in comparison
to those currently available.
The original discovery synthesis8 was based on a Mu¨nchnone
cyclization9 of R-amino acid 2(Scheme 1), but imidazole 3 was
formed as the minor regioisomer and proved challenging to
isolate cleanly by chromatography. An alternate preparation
from R-amino ꢀ-ketoester 5 also afforded low conversion to 3
but avoided the formation of regiochemical isomers. Thus, 3
could be readily isolated from other byproduct via flash
chromatography, which enabled the second-generation synthesis
(Scheme 2) capable of providing gram quantities of 1 for
preclinical toxicological studies. However, it became evident
that several steps of the process would need to be significantly
improved to provide larger quantities of API.
Acylation of glycine diphenylimine 7 with isobutyryl
chloride originally required three cryogenic reactors and a
tedious workup. Amine salt 8 was typically isolated by
crystallization from water, making the subsequent moisture-
sensitive amide formation extremely difficult. To make matters
worse, 8 was prone to decomposition upon vacuum drying
When analyzing the existing synthesis of 1 for potential
process improvements, it was determined that identification of
an efficient synthesis of the key imidazole core was paramount.
* To whom correspondence should be addressed: E-mail: dan.bowles@
pfizer.com. Telephone: (860)-686-0925
(1) Tiwari, A.; Bansal, V.; Chugh, A.; Mookhtiar, K Expert Opin. Drug
Saf. 2006, 5, 651–666.
(2) Grundy, S. M. J. Intern. Med. 1997, 241, 295–306.
(3) Austin, M. A.; Hutter, C. M.; Zimmern, R. L.; Humphries, S. E. Am. J.
Epidemiol. 2004, 5, 421–429.
(4) Jones, P.; Kafonek, S.; Laurora, I.; Hunninghake, D. Am. J. Cardiol.
1998, 81 (5), 582–587.
(5) Cerivastatin was withdrawn from the world market because of its
potential for severe myotoxic effects. For a review, see: Rosenson, R.
Am. J. Med. 2004, 116, 408–416.
(8) Bolton, G. L.; Bowles, D. M.; Boyles, D. C.; Howard, W. A., Jr.;
Hutchings, R. H.; Kennedy, R. M.; Park, W. K. C.; Poel, T. J.; Song,
Y. Preparation of Novel Imidazoles as HMG Co-A Reductase
Inhibitors for Use in Treating Hyperlipidemia and Other Diseases. U.S.
Pat. Appl. 0239857 A1, 2005.
(9) Padwa, A.; Gingrich, H. L.; Lim, R. Tetrahedron Lett. 1980, 21 (36),
3419–3422.
(6) Graham, D. J.; Staffa, J. A.; Shatin, D.; Rade, S. E.; Schech, S. D.;
La Grenade, L.; Gurwitz, J. H.; Chan, K. A.; Goodman, M. J.; Platt,
R. JAMA, J. Am. Med. Assoc. 2004, 292, 2585–2590.
(7) Hamelin, B. A.; Turgeon, J. Trends Pharmacol. Sci. 1998, 19, 26–
37.
10.1021/op800092e CCC: $40.75
Published on Web 09/27/2008
2008 American Chemical Society
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1183

Scheme 2. Second-generation synthesis of 1a
Scheme 3. Optimized second-generation synthesis of 1a
a Reagents and conditions: (a) benzophenone imine, CH2Cl2, 20 °C, 8 h, 99%;
(b) i. KOtBu, isobutyryl chloride, THF, -78 °C, 30-70 %; ii. HCl (aq); (c)
4-fluorobenzoyl chloride, Et3N, CH2Cl2, 0 °C, 30 min, 45-85%; (d) xylene,
AcOH, TsOH (cat.), reflux, 16 h, 20%; (e) Pd/C, H2, 50 psi, 25 °C, 4 h, 91%;
(f) benzylamine CDI, DMF, 0 °C to 20 °C, 2 h, 90%; (g) TFA, CH2Cl2, 0 °C,
30 min, 52%; (h) NaOH (aq), THF, 50 °C, 2 h, 92%; * step required
chromatography.
a Reagents and conditions: (a) i. KOtBu, isobutyryl chloride, THF, -70 °C;
ii. HCl (aq); (b) K2CO (aq), CH2CL2, 4-fluorobenzoylchloride, 0 °C, 0.5 h, 75%
(2 steps); (c) benzylamine, Et3N, heptane 70 °C, 6 h, 80%; (d) 4, benzoic acid,
heptane, 82 °C, 48 h; (e) MeOH, HCl (aq), 45 °C, 12 h; (f) i. NaOH (aq), IPA,
95 °C, 6 h; ii. HCl (aq) to pH ) 4, MTBE, 50 °C, 1 h; iii. NaOH (aq), IPA; (g)
IPA (aq), reflux to 18 °C, 57% from 12; [] not isolated.
above 35 °C, so the overall conversion of imine 7 to ketoamide
5 was not robust. Cyclization of 5 with amine 410 afforded 3,
but this reaction suffered from a variety of side reactions,
resulting in a poor overall yield (<20%). Debenzylation and
CDI coupling afforded smooth conversion to the amide 10, but
acetonide deprotection with trifluoroacetic acid resulted in
formation of a variety of impurities stemming from ꢀ-hydroxy
elimination. Sodium salt 1 was isolated as a mixture of
amorphous and crystalline solids, and excess sodium hydroxide
contaminated the final product. A total of five chromatography
steps were also necessary via the original synthetic pathway.
In this paper we report a series of practical refinements of
the second-generation synthesis of 1 that proved more amenable
to preparative scale-up. The acylating, amide-forming, and
imidazole-forming steps were all significantly improved and
simplified from the original procedure. An improved protocol
for isolation of crystalline 1 allowed all chromatography steps
to be removed from the synthesis.
however, proved to be difficult. Acidic workup of the reaction
mixture obtained using the second-generation chemistry resulted
in isolation of approximately 20% of hydrolyzed glycine ester
6 as an unwanted byproduct. However simply charging 1.0
equiv of both reagents led to the optimized results: a 93:5 ratio
of 8 to 6 in 83% yield. In this procedure, imine 7 was added
slowly to a -70 °C slurry potassium tert-butoxide in THF, and
the resulting anion was transferred into a solution of isobutyryl
chloride. Acylated intermediate 14 was hydrolyzed with aqueous
HCl, and benzophenone was removed by extraction with
MTBE. Amine salt 8 was either crystallized from the aqueous
layer in 83% yield (99% (area %) HPLC purity) or was more
preferably carried forward as a crude aqueous solution.
The problematic anhydrous amidation conditions were
replaced with a more effective Scho¨tten-Baumann procedure
in which the aqueous solution of 8 was treated with potassium
carbonate and dichloromethane, followed by slow addition of
4-fluorobenzoyl chloride. This made it unnecessary to dry and
isolate 8 before amidation, as ketoamide 5 was isolated by
crystallization from MTBE/heptane in 90% yield (>98% HPLC
purity).
The second-generation synthesis (Scheme 2) had required a
three-step sequence including (i) cyclization to afford imidazole
3, (ii) debenzylation to produce acid 9, and (iii) CDI coupling
with benzylamine to produce amide 10. However, the benzyl
ester in ketoamide 5 was prone to trans-amidation with amine
4 during the formation of 3, resulting in a poor yield (<20%)
for this key step. We reasoned that installation of the desired
benzyl amide prior to the cyclization might “block” this site to
the undesired side reaction, potentially improving the formation
Results and Discussion
The second-generation route was modified to enable the
synthesis of several hundred grams of 1, in 16% yield over
seven steps (Scheme 3). The preparation of masked glycine 7
was not altered from the original procedure and was readily
isolated from the imine exchange reaction of HCl salt 611 with
benzophenone imine. The subsequent acylation of imine 7,
(10) Baumann, K. L.; Butler, D. E.; Deering, C. F.; Mennen, K. E.; Millar,
A.; Nanninga, T. N.; Palmer, C. W.; Roth, B. D. Tetrahedron Lett.
1992, 33 (17), 2283–2284.
(11) The corresponding tosylate and mesylate salts, ethyl and methyl esters,
and benzyl amide were also investigated.
1184
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

Scheme 4. Retro-Claisen dealkylation of ketoamides 5 and
12
Scheme 5. Lactone 13, a useful intermediate toward the
purification of 1a
a Reagents and conditions: (a) HCl (aq), 50 °C, 1 h; (b) NaOH (aq), IPA; (c)
98:2 IPA/water, 80 °C to 18 °C.
Despite these improvements, it seemed that flash chroma-
tography would have still been required to isolate imidazole
10 in high purity. Instead, as a general practice this intermediate
was carried forward as a crude concentrate (60-75% HPLC
purity), and critical impurities were removed downstream.
Substitution of methanolic HCl (aq) for TFA resulted in smooth
conversion of acetonide 10 to diol 11, while retaining the chiral
integrity.13 The tert-butyl ester was cleaved with excess sodium
hydroxide, and free acid 19 (Scheme 5) was extracted into
MTBE following pH adjustment with HCl. The organic solvents
were replaced with 2-propanol via vacuum distillation to
produce a 15:1 ratio of 19 (free acid) to lactone 13 (Scheme
5).14 As a result, the end point of the subsequent sodium
hydroxide addition (1.0 molar equiv) was easily reached by
slowly metering in caustic until the level of 13 fell to <1% by
HPLC analysis.15 Residual water was removed to 2%16 via
azeotropic distillation, and sodium salt 1 was isolated by
filtration.
of the imidazole. However, the desired trans-amidation resulting
from treatment of 5 with benzylamine resulted primarily in retro-
Claisen dealkylation (Scheme 4) to produce 18.
The conversion of 5 to 12 (Scheme 3) was further examined
in a variety of solvents. Highly polar solvents such as DMF,
DMSO and benzylamine resulted in complete consumption of
5 but exacerbated the decomposition. Nonpolar solvents such
as hexane, heptane and methyl cyclohexane appeared to
minimize formation of 18, but conversion often stalled due to
poor solubility of both 5 and 12. Use of non-nucleophilic amines
such as Hunig’s base and triethylamine as cosolvents improved
solubility and conversion without significantly increasing
dealkylation. Under the most effective conditions screened, 5
was treated with 2 equiv of benzylamine in 2:1 heptane/
triethylamine to afford 12 as a fluffy solid that was easily filtered
away from the crude reaction mixture in good purity.
With 12 in hand, our focus turned toward optimization of
the original condensation conditions (refluxing xylenes, acetic
acid, 72 h). A variety of solvent and acid combinations were
studied. Installation of the amide functionality in 12, as
predicted, immediately improved the isolated yield of 1012
(Scheme 3) in the cyclization from 20% to 41%. Several
nonpolar solvents performed well, and heptane was especially
effective. Most polar solvents afforded poor conversion to
desired product.
During the course of this study, it was determined that a
mild acid additive was crucial for stabilization of amine 4
against dimerization. We discovered that under the original
cyclization conditions, acetic acid was inadvertently removed
via its azeotrope with heptane and was trapped in the aqueous
layer in the Dean-Stark apparatus, resulting in poor conversion
to 10. Cyclization in neat acetic acid failed. A variety of
carboxylic acids with increased resistance to azeotropic distil-
lation, including pivalic, benzoic, tosic, and solid-supported acid
resins were screened, and benzoic acid, which was easily
washed away afterward, enabled significant improvement. As
an added benefit, replacement of the original reagents xylenes
and acetic acid with heptane and benzoic acid, respectively,
allowed the necessary equivalents of amine 4 to be optimized
from 5 down to 1.1. The isolated yield of 10 was improved to
72%, an increase of >50% over the original procedure with
benzyl ester 5.
Conclusions
A low yield second-generation synthetic route was trans-
formed into a synthesis capable of manufacturing multiple
kilograms of API within a scale-up facility, provided cryogenic
capability is available. The acylating and amide-forming steps
were combined into a one-pot procedure, resulting in improved
robustness and yield. Replacing the benzyl ester of 5 with the
corresponding benzyl amide increased the stability of down-
stream intermediates. The key imidazole-forming step was
optimized to improve the yield from 20% to >70%, and the
number of equivalents of amine 4 was decreased from 5 to 1.1.
All five chromatographic purification steps were removed,
enabling increased throughput and decreased processing time.
Overall, the synthesis of imidazole 1 was improved from 4%
to 16% yield over seven steps and was isolated in >99% HPLC
purity (>99.5% de).17
Experimental Section
All reagents were purchased from commercial suppliers and
used as received unless otherwise specified. All reactions were
(13) No ꢀ-elimination observed under these conditions.
(14) Lactone 13 was not crystallized despite repeated efforts.
(15) Sodium hydroxide selectively consumes free acid 19, followed by
reaction with lactone 13, in that order. This phenomenon allowed
careful titration to the desired addition endpoint.
(16) Determined by Karl Fischer analysis.
(17) A laboratory-scale batch was carried forward from raw materials to
>250 g of 1 within 2 weeks time. Nearly 600 g of non-cGMP API
was prepared by this method.
(12) Isolated by column chromatography in >98% HPLC purity for this
study.
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1185

performed under a dry nitrogen atmosphere. NMR spectra were
dichloromethane (1.0 L). The mixture was cooled to between
-10 and 0 °C, and a solution of potassium carbonate (546.0 g,
3.95 mol) in water (1.5 L) was added at a rate of 2.5 mL/min,
maintaining the reaction temperature below 5 °C. The reaction
was recooled to 0 °C and was charged with a solution of
4-fluorobenzoyl chloride (209 mL, 1.74 mol) in dichlo-
romethane (500 mL) at a rate of 1.5 mL/min, maintaining the
reaction temperature below 5 °C. The reaction layers were
separated after 30 min, and the aqueous portion was extracted
with dichloromethane (500 mL). The combined organic layers
were washed sequentially with HCl (0.1 M, 500 mL) and water
(2 L) and then concentrated under vacuum to produce a yellow
solid that was reconcentrated18 from MTBE (0.5 L) and
crystallized from a mixture of refluxing MTBE (1 L) and
heptane (2.5 L). The resulting solid was collected by filtration
and washed with heptane (0.5 L) and then dried for 12 h at 35
°C under vacuum to afford 5 (504.0 g, 90%) as a fluffy white
solid. 1H NMR (CDCl3): δ 7.81 (dd, J ) 7.0 and 4.8 Hz, 2H),
7.38-7.29 (m, 5H), 7.09 (dd, J ) 8.5 and 8.6 Hz, 2H), 5.60
(d, J ) 6.5 Hz, 1H), 5.22 (dd, J ) 21.2 and 12.2 Hz, 2H),
3.07-3.02 (m, 1H), 1.20 (d, J ) 7.0 Hz, 3H), 1.00 (d, J ) 7.0
Hz, 3H). 19F NMR (CDCl3): δ -107.54. HPLC (area %) purity:
99.2%. MS m/z calcd for C20H20FNO4 357.14, found 358.36
(M+).
1
measured at 400 and 376 MHz, for H and 19F, respectively.
HPLC analyses of chemical purity were carried out using a
YMC Pack Pro C18 column (150 mm × 4.6 mm, 3 µm) with
0.2% HClO4 in 95:5 water/acetonitrile (mobile phase A) and
acetonitrile (mobile phase B) with a gradient method: 20%
eluent B linearly increased to 95% eluent B over 20 min, 1.0
mL/min, 235 nm. Chiral purity analyses were carried out using
a Daicel Chiralpak AD column (250 mm × 4.6 mm, 10 µm)
with an isocratic mobile phase of hexanes/2-propanol, 85:15;
1.0 mL/min, 210 nm.
(Benzhydrylidene-amino)acetic Acid Benzyl Ester (7). A
5-L round-bottom flask equipped with an overhead stirrer and
nitrogen inlet was charged with glycine benzyl ester hydro-
chloride 6 (505.2 g, 2.51 mol) and CH2Cl2 (3.0 L). To this
mixture was charged benzophenone imine (471.1 g, 2.60 mol)
in one portion. The reaction was stirred at 20 °C for 8 h until
TLC analysis (1:1 ethyl acetate/heptane) confirmed full con-
sumption of 6. The precipitated ammonium salt was removed
with a short pad of Celite, and the filter cake was rinsed with
CH2Cl2 (0.5 L). The filtrates were concentrated under reduced
pressure, and the resulting solid was dried for 12 h at 35 °C
under vacuum to afford 7 (837.3 g, 101%) as a white solid that
was carried on without further purification. 1H NMR(DMSO-
d6): δ 7.53-7.25 (m, 13H), 7.12 (m, 2H), 5.10 (s, 2H), and
4.17 (s, 2H). MS m/z calcd for C22H19NO2 329.14, found 330.36
(M+).
2-Amino-4-methyl-3-oxo-pentanoic Acid Benzyl Ester
Hydrochloride (8). A 3-L round-bottom flask equipped with
an overhead stirrer, cooling bath, and nitrogen inlet was charged
with a slurry of potassium tert-butoxide (120.0 g, 1.07 mol) in
THF (600 mL). The slurry was cooled to below -70 °C and
was charged with a solution of 7 (350.3 g, 1.06 mol) in THF
(800 mL) over 0.5 h, maintaining the reaction temperature
below -40 °C. The reaction mixture was recooled to -70 °C
and was then transferred into a second 5 L reactor containing
a -70 °C solution of isobutyryl chloride (114.0 g, 112.1 mL,
1.07 equiv) in THF (300 mL), maintaining the reaction
temperature in the receiving reactor below -55 °C. The reaction
was stirred for 3 h between -55 and -65 °C and was then
removed from the cold bath and quenched with HCl (3 M, 3.18
mol) in one portion. The reaction was allowed to warm to -10
°C, and was diluted with water (1 L) followed by extraction
with MTBE (2 × 1 L). The aqueous layer was separated and
distilled at 40 °C for 1 h under vacuum (60 Torr) to remove
any residual MTBE. The resulting aqueous mixture was placed
in the refrigerator between 0 and 3 °C for 12 h and was filtered
to produce 8 (238.5 g, 83%) as a white solid that was carried
forward without additional drying. Alternatively, crude 8 can
be carried forward without crystallization. 1H NMR (CD3OD):
δ 7.37-7.30 (m, 5H), 5.29-5.18 (dd, J ) 23.8 and 12.2 Hz,
2H), 3.16-3.00 (m, 2H), 1.13 (d, J ) 7.1 Hz, 3H), 1.00 (d, J
) 6.9 Hz, 3H). MS m/z calcd for C13H18NO3Cl 271.10, found
236.31 (M+ less HCl).
N-(1-Benzylcarbamoyl-3-methyl-2-oxo-butyl)-4-fluoroben-
zamide (12). A 3-L round-bottom flask equipped with an
overhead stirrer, heating mantle, reflux condenser, and nitrogen
inlet was charged with a slurry of 5 (500.0 g, 1.4 mol) in heptane
(1.5 L) and triethylamine (750 mL). Benzylamine (375.0 g, 3.5
mol) was added in one portion, and the resulting slurry was
heated to 70 °C for 6 h. The reaction was cooled to 18 °C and
was filtered to produce a wet cake that was recrystallized from
refluxing heptane (1.5 L). The resulting solids were collected
by filtration and dried at 55 °C under vacuum for 12 h to afford
1
12 (401.2 g, 80%) as a pale-yellow, fluffy solid. H NMR
(CDCl3): δ 7.85-7.80 (m, 2H), 7.70, (d, J ) 6.5 Hz, 1H),
7.41-7.10 (m, 7H), 5.33 (d, J ) 7.0 Hz, 1H), 4.42 (m, 2H),
3.17-3.14 (m, 1H), and 1.10 (m, 6H). 19F NMR (CDCl3): δ
-106.95. MS m/z calcd for C20H21FN2O3 356.15, found 357.12
(M+).
7-[4-Benzylcarbamoyl-2-(4-fluorophenyl)-5-isopropyl-
imidazol-1-yl]-3,5-dihydroxyheptanoic Acid, Sodium Salt
(1). Cyclization. A 5-L round-bottom flask outfitted with an
overhead stirrer and Dean-Stark apparatus was charged
sequentially with benzyl amide 12 (356.7 g, 1.0 mol), amine 4
(183.0 g, 1.5 mol), benzoic acid (183.0 g, 1.5 mol), and heptane
(3 L). The resulting slurry was heated to reflux for 18 h, and
any water generated during the course of the reaction was
removed via the Dean-Stark apparatus. The reaction progres-
sion was monitored by water collection and by HPLC analysis
over 48 h, until less than 0.5% of 12 remained. The reactor
was refitted with a distillation head and a total of 1.8 L of
heptane was distilled out under ambient pressure. The reactor
was cooled to 40 °C and charged with MTBE (1 L). The
resulting solution was extracted with saturated sodium carbonate
2-(4-Fluorobenzoylamino)-4-methyl-3-oxo-pentanoic Acid
Benzyl Ester (5). A 5-L round-bottom flask equipped with an
overhead stirrer, cooling bath, addition funnel, and nitrogen inlet
was charged with 8 (427.8 g, 1.57 mol), water (1.0 L) and
(18) Azeotropic drying.
1186
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

solution (2 × 750 mL) followed by HCl (0.5 N, 1 L). Crude
10 was carried forward in the organic phase without further
purification.
Acetonide RemoVal. The organic layer was returned to the
reactor and was diluted with methyl alcohol (1.8 L) and a
solution of HCl (12 N, 34.0 g) in water (1.3 L). The mixture
was heated to 45 °C for 3.5 h, until acetonide 10 was fully
consumed by HPLC analysis. A crude solution of diol 11 and
lactone 13 was carried forward into the next step.
tert-Butyl Ester CleaVage. The previous reaction mixture was
charged with aqueous sodium hydroxide (50 wt %/wt) to pH
) 13,19 and was heated to 50 °C for 2 h, until tert-butyl ester
11 was determined to be fully consumed by HPLC analysis.
The reaction was cooled to 35 °C, diluted with water (1 L),
and extracted with MTBE (2 × 1 L). The aqueous product-
containing layer was distilled under ambient pressure to a
reaction temperature of 97 °C, and the resulting solution was
cooled to 50 °C and held at this temperature while adding HCl
(6 N) until the pH reached 7. The reactor was charged with
MTBE (1 L) and the pH of the mixture was further adjusted to
pH 4 ( 0.5 with HCl (1 N). The product-containing organic
layer20 was distilled under ambient pressure down to a volume
of 0.6 L, and 2-propanol (1.5 L) was added. Distillation was
continued, refilling with additional 2-propanol as necessary, until
the pot temperature reached 82 °C, and the resulting crude
mixture of free acid 19 and lactone 13 was distilled to a reaction
volume of 0.5 L.
aqueous 2-propanol as necessary, and the reactor was cooled
at a rate of 5 °C/h to 20 °C. The resulting slurry was stirred for
an additional 3 h, and the solids were collected by filtration
and rinsed with anhydrous 2-propanol (300 mL) to produce
crude 1 (92% (area %) HPLC purity) that was carried forward
into the recrystallization.
Recrystallization. Sodium salt 1 was redissolved into 80:20
2-propanol/water (700 mL) at reflux under a distillation head,
and 2-propanol (500 mL) was added in one portion. The mixture
was reheated to reflux, and additional 2-propanol was added,
as necessary, while continuously distilling, until Karl Fischer
analysis indicated between 1-2% of water (wt %) remained.
The reactor was cooled at a rate of 5 °C/h to 20 °C, and the
resulting slurry was stirred for 3 h and filtered. The solid product
cake was rinsed with anhydrous 2-propanol (300 mL) and dried
at 75 °C under vacuum for 12 h to afford 294.8 g (57%, >
99% (area %) HPLC chemical purity, > 99.5% de) of 1 as a
1
crystalline white solid. H NMR (CD3OD): δ 7.62 (m, 2H),
7.33-7.19 (m, 7H), 4.95 (bs, 1H), 4.51 (s, 2H), 4.21 (ddd, J )
14.85, 11.33 and 5.08 Hz, 1H), 4.11-3.92 (m, 2H), 3.73 (m,
1H), 3.46 (m, 1H), 2.29 (dd, J ) 15.24 and 5.47 Hz, 1H), 2.23
(dd, J ) 15.04 and 7.42, Hz, 1H), 1.89-1.76 (m, 1H),
1.73-1.61 (m, 1H), 1.61-1.54 (m, 1H), 1.48 (m, 7H). 19F NMR
(CD3OD): δ -113.88. Anal. Calculated for C27H31F1N3NaO5:
C, 62.42; H, 6.01; N, 8.09; Na, 4.40. Found: C, 62.32; H, 5.93;
N, 8.05; Na, 4.39. Karl Fischer: 0.36% water. IR(neat) νmax
)
1657, 1574, 1512, 1411, 1223, 846, and 700 cm-1.
Salt Formation/Isolation. Aqueous sodium hydroxide (50
wt %/wt) was added to the reaction mixture in portions, and
the level of lactone 13 was carefully monitored by HPLC
analysis. Addition of sodium hydroxide was halted when only
0.2 to 0.5% of 13 remained after 10 min of stirring.21 Additional
anhydrous isopropyl alcohol (0.5 L) was added, the distillation
head was refitted, and the solution was distilled under ambient
pressure until Karl Fischer analysis indicated between 1 and
2% water (solution wt %) remaining.22 The reaction volume
was readjusted to a total volume of 1.0 L with additional 2%
Acknowledgment
We thank Mike Lovdahl and Eric Nord for analytical
assistance, and Dan Belmont, Jerry Clark, David Erdman,
Randy DeJong, Michael Pamment, Derek Pflum, James Saenz,
Michael Stier, Bruce Roth, and Brad Tait for helpful suggestions.
Supporting Information Available
Tables showing solvent effects on formation of amide 12
and influence of solvent and acid on yield of imidazole 10. This
information is available free of charge via the Internet at http:
//pubs.acs.org.
(19) Additional sodium hydroxide was added, as necessary, to maintain
the pH at 13 until this step was completed.
(20) As much as 10% of 19 was converted to lactone 13 during this pH
adjustment.
(21) Any unreacted 13 was selectively purged in the subsequent crystal-
lization step.
(22) Additional anhydrous isopropyl alcohol was added, as necessary, to
maintain the reaction volume near 1 L.
Received for review April 18, 2008.
OP800092E
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1187