Pilot-plant preparation of an αvβ3 integrin antagonist. Part 3. Process research and development of a diisopropylcarbodiimide and catalytic 1-hydroxybenzotriazole peptide coupling

Article abstract of DOI:10.1021/op900173f

Studies directed toward the process research, development, and scale-up preparation of the potential α_vβ₃ integrin antagonist 1 are described. A convergent approach is detailed wherein tetrahydropyrimidine hydroxybenzoic acid 2 is li

Full text of DOI:10.1021/op900173f

Organic Process Research & Development 2009, 13, 1088–1093
Pilot-Plant Preparation of an rvꢀ3 Integrin Antagonist. Part 3. Process Research
and Development of a Diisopropylcarbodiimide and Catalytic 1-Hydroxybenzotriazole
Peptide Coupling†
Jerry D. Clark,‡,* D. Keith Anderson,‡ David V. Banaszak,⊥ Derek B. Brown,⊥ Ann M. Czyzewski,⊥ Albert D. Edeny,⊥
Puneh S. Forouzi,⊥ Donald J. Gallagher,⊥ V. H. Iskos,⊥ H. Peter Kleine,‡ Carl M. Knable,⊥ Melissa K. Lantz,⊥
Mark A. Lapack,⊥ Christine M. V. Moore,⊥ Frank W. Muellner,⊥ James B. Murphy,⊥ Carlos A. Orihuela,⊥ Mark A. Pietz,⊥
Thomas E. Rogers,⊥ Peter G. Ruminski,‡ Harish K. Santhanam,⊥ Tobin C. Schilke,⊥ Ajit S. Shah,⊥ Ahmad Y. Sheikh,⊥
Gerald A. Weisenburger,§ and Bruce E. Wise⊥
Pfizer Global Research and DeVelopment, St. Louis Laboratories, 700 Chesterfield Parkway West, Chesterfield, Missouri 63017,
and Pfizer Global Research and DeVelopment, Eastern Point Road, Groton, Connecticut 06340
Abstract:
Studies directed toward the process research, development,
and scale-up preparation of the potential rvꢀ3 integrin
antagonist 1 are described. A convergent approach is
detailed wherein tetrahydropyrimidine hydroxybenzoic
acid 2 is linked to the ꢀ-amino acid ester 3 via a
diisoproylcarbodiimide, catalytic HOBt coupling reac-
tion. Saponification of the resulting ethyl ester, isolation
of the corresponding zwitterion, H3PO4 salt formation,
crystallization, and acetonitrile-to-acetone exchange
give rise to 1 as a crystalline monohydrate in 67%
overall yield from 4.
Results and Discussion
Preparation of 1 began by converting 44 to 3. Initial
laboratory-scale experiments facilitated the desired deprotection
through the use of 4 N HCl in dioxane, but upon further
investigation, use of this reagent on scale gave rise to 3 that
Introduction
The integrin Rvꢀ3 plays an essential role in angiogen-
esis, the process by which new blood vessels form from
pre-existing blood vessels.1 Angiogenesis is required for
tumor growth, and therefore, antagonists of Rvꢀ3 are being
studied for the treatment of cancer. In a program directed
toward the discovery of such antagonist, an improved
synthesis of the monohydrate, monophosphoric acid salt
1 was required.2
(1) (a) Liu, Z.; Wang, F.; Chen, X. Drug DeV. Res. 2008, 69, 329. (b)
Roberto, R.; Angelo, V.; Domenico, R.; Francesco, D. R.; Francesca,
M.; Franco, D. Haematologica 2002, 87, 836. (c) Ruminski, P. G.;
Rogers, T. E.; Rico, J. E.; Nickols, G. A.; Westlin, W. F.; Engleman,
V. W.; Settle, S. L.; Keene, J. L.; Freeman, S. K.; Carron, C. P.; Meyer,
D. M. 218th National Meeting of the American Chemical Society,
New Orleans, 1999; Book of Abstracts; American Chemical Society:
Washington, DC, 1999. (d) Ruminski, P. G.; Rogers, T. E.; Rico, J. E.;
Nickols, G. A.; Engleman, V. W.; Settle, S. L.; Keene, J. L.; Freeman,
S. K.; Carron, C. P.; Meyer, D. M. 217th National Meeting of the
American Chemical Society, Anaheim, CA, 1999; Book of Abstracts;
American Chemical Society: Washington, DC, 1999. (e) Flynn, D. L.;
Abood, N.; Bovy, P.; Garland, R.; Hockerman, S.; Lindmark, R. J.;
Schretzman, L.; Rico, J.; Rogers, T. 210th National Meeting of the
American Chemical Society, Chicago, IL, 1995; Book of Abstracts;
American Chemical Society: Washington, DC, 1995.
(2) (a) Collins, J. T.; Devadas, B.; Lu, H.-F.; Malecha, J. W.; Miyashiro,
J. M.; Nagarajan, S.; Rico, J. G.; Rogers, T. E. U.S. Patent 6,100,423,
2000. CAS 1098091-47-4. (b) Malecha, J. W.; Fraher, T. P. U.S. Patent
6,172,256, 2001. CAS 253866-45-5. (c) Colson, P.-J.; Awasthi, A. K.;
Nagarajan, S. R. U.S. Patent 6,414,180, 2002; CAS 290826-60-7.
(3) (a) Clark, J. D.; Collins, J. T.; Kleine, H. P.; Weisenburger, G. A.;
Anderson, D. K. Org. Process Res. DeV. 2004, 8 (4), 571. (b) Clark,
J. D.; Weisenburger, G. A.; Anderson, D. K.; Colson, P.-J.; Edney,
A. D.; Gallagher, D. J.; Kleine, H. P.; Knable, C. M.; Lantz, M. K.;
Moore, C. M. V.; Murphy, J. B.; Rogers, T. E.; Ruminski, P. G.; Shah,
A. S.; Storer, N.; Wise, B. E. Org. Process Res. DeV. 2004, 8 (1), 51.
(c) Clark, J. D.; Anderson, D. K.; Collins, J. T.; Colson, P.-J.; Edney,
A. D.; Gallagher, D. J.; Kleine, H. P.; Knable, C. M.; Lantz, M. K.;
Moore, C. M. V.; Ruminski, P. G.; Weisenburger, G. A.; Wise, B. E.
National Organic Chemistry Symposium, Indiana University, Bloom-
ington, IN, 2003. (d) Clark, J. D. Gordon Research Conference,
Organic Reactions and Processes, Bristol, RI, July 23, 2002.
Envisioned was a convergent synthesis wherein hydroxy-
benzoic acid 2 would be linked to the ꢀ-amino acid amide
3. As part of a series of reports,3 herein we describe the
chemical process research, development, and large-scale
preparation of 1.
† Manuscript dedicated to Professor Paul A. Grieco on the occasion of his
65th birthday.
* To whom correspondence should be addressed. E-mail: jerry.d.clark@
pfizer.com.
‡ Pfizer Global Research and Development, St. Louis Laboratories.
§ Pfizer Global Research and Development, Groton.
⊥ No longer employed by Pfizer.
(4) Weisenburger, G. A.; Anderson, D. K.; Clark, J. D.; Edney, A. D.;
Karbin, P. S.; Gallagher, D. J.; Knable, C. M.; Pietz, M. A. Org.
Process Res. DeV. 2009, 13 (1), 60.
1088
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development
10.1021/op900173f CCC: $40.75  2009 American Chemical Society
Published on Web 09/04/2009

Table 1. Summary of the thermokinetic data for the
conversion of 4 to 3
total heat
[kJ]
intrinsic heat
[kJ/mol x]
adiabatic temp.
unit operation
rise [°C]
hydrogen chloride -128
136/mol x ) AcCl
+79
preparation
preparation of 3
-0.91 9.2/mol x ) 4
+0.46
was hygroscopic and contained several percent of the corre-
sponding, undesired carboxylic acid. These findings in conjuga-
tion with the potential safety issues associated with the use of
4 N HCl in dioxane on scale prompted development of another
procedure.
After some experimentation it was determined that depro-
tection and isolation of 3 could be successfully obtained by
generating anhydrous HCl in situ from the reaction of 10 equiv
of acetyl chloride and ethanol at e10 °C. Following workup, 3
was isolated by filtration as a white solid in >95% yield and
>96 HPLC % purity.
In order to facilitate the technology transfer of this reaction
to the pilot plant and to ascertain the thermal and safety
characteristics of the chemistry, reaction calorimetry experiments
as well as IR and MS head space gas analysis for the conversion
were performed.5 The results of the calorimetry experiments
are provided in Table 1. Both reactions produced moderate to
low levels of heat, demonstrating that the chemistry should be
easily controlled in a standard reactor setup.
The major gaseous products detected in the reactor headspace
were ethanol, CO2, HCl, isobutylene, tert-butyl chloride, ethyl
acetate, and acetyl chloride.6 Only 43% or 0.04 mol of the
theoretical quantity of 0.0938 mol of CO2 were measured. This
was most likely due to the solubility of CO2 in ethanol7 and,
perhaps, losses in the reactor configuration. Also detected was
0.2 mmol of isobutylene. This translated to only 0.21% of the
available moles of substrate vented as isobutylene. It was
therefore determined that a standard scrubber configuration
would support the disposition of the gaseous products.
The aforementioned procedure was practical for a laboratory-
scale reaction, but further work on the concentration and isola-
tion of the product on large scale was required. In this regard,
the solubility of 3 in ethyl acetate/ethanol solvent mixtures was
measured. The level of ethanol in the solvent mixture was an
important parameter for the efficient isolation of 3. As illustrated
in Figure 1, the solubility of 3 in ethyl acetate/ethanol increased
significantly as the wt % of ethanol increases versus ethyl
acetate. This was verified experimentally. For example, a 30%
yield loss of 3 was realized if the ethanol content was 12 wt
%. At 5 wt % of the solvent, 12% of 3 was lost.
Figure 1. Solubility of 3 in ethanol/ethyl acetate.
equiv of acetyl chloride at e10 °C. The solvent was then
exchanged to ethyl acetate, being sure to decrease the ethanol
content to <1 wt %. The product was collected, washed with
cold ethyl acetate, and blown partially dry with warm N2.
Drying was completed in a tumble dryer affording 125 kg of 3
as a white solid in 93-95% yield.
Coupling Process Development. Process development
toward the coupling of 2 with 3 to prepare the ethyl ester of 1
began by surveying a number of coupling reagents.8 1,3-
Diisopropylcarbodiimide (DIC), 1,3-dicyclohexylcarbodiimide
(DCC), 2-chloro-4,6-dimethoxy-1,3,5-triazine (CDMT),9 and
benzotriazol-1-yloxytris(dimethylamino) phosphonium hexaflu-
orophosphate (BOP) were all effective coupling agents. Table
2 summarizes these results.10
Of these coupling reagents, DIC was employed in all scale-
up work. This reagent proved easy to source, was cost-effective
and known to provide greater operator safety.11 Preliminary
research determined that 1.3 equiv of DIC at 20 °C for 23 h
provided the best reaction performance. The amount and number
of reaction impurities increased when either >1.3 equiv of DIC
was employed or the reaction was carried out at >30 °C.
The best solvent system for the desired coupling was found
to be a 1:1 (v:v) mixture of DMF/CH2Cl2.12 Other aprotic
solvents as well as alcohols facilitated the desired coupling
reaction. The reaction was faster in alcohols versus DMF/
CH2Cl2 but required the addition of 1 equiv of HCl, and more
importantly, the workup procedure from alcohols proved more
difficult. The rate of the coupling reaction in NMP was greater
(8) Bodanszky, M. Peptide Chemistry, 2nd ed.; Springer-Verlag: New
York, 1993.
(9) Formation of the CDMT adducts of 2 were observed as well as the
desired coupling product. (a) Clark, J. D. Midwest Pharmaceutical
Chemistry Consortium Symposium, Kalamazoo, MI, October 5, 2001.
(b) Clark, J. D. 15th Annual Organic Chemistry Day, University of
Missouri, Columbia, MO, April 27, 2002.
(10) N,N′-Carbonyldiimidazole, 1-(3-dimethylaminopropyl)-3-ethylcarbo-
diimide HCl, bis(2-oxo-3-oxazolidinyl)phosphinic chloride, 2-chloro-
1-methylpyridinium iodide, diethyl chlorophosphate, trimethylacetyl
chloride, isovaleryl chloride, isobutyryl chloride, 1-propanephosphonic
acid cyclic anhydride, 2-nitrophenol, 4-nitrophenol, 2,4,5-trichlorophe-
nol, and O-benzotriazolyl-N,N,N′,N′-tetramethyluronium hexafluoro-
phosphate were ineffective coupling agents when using 2 equiv relative
to 2.
With this data in hand, 3 was successfully prepared in two
pilot plants. A solution of 4 in ethanol was charged with 10
(5) The calorimeter used in this study was a RC1 SV01 1L reactor
equipped with a pitched blade impeller, addition funnel, a thermo-
couple, calibration heater, a P6890 MSD GC/MS system used for
sampling from the RC1 SV01 reactor head space, and a ASI
ReactIR1000 with a diamond ATR probe. IR spectrum analysis was
completed using ASI Concert IR analysis software.
(6) Except for CO2 and isobutylene, all gas flows were calculated by
assuming that the respond factor of gas/Ar was equal to 1. Quantitative
CO2 and isobutylene data were calculated by the response factor of
CO2 (m/z 44)/Ar (m/z 40) ) 0.74, isobutylene (m/z 56)/Ar (m/z 40) )
0.15.
(11) Hayes, B. B.; Gerber, P. C.; Griffey, S. S.; Meade, B. J. Drug Chem.
Toxicol. 1998, 24 (2), 195.
(7) Dias, E. L.; Hettenbach, K. W.; Am Ende, D. J. Org. Process Res.
DeV. 2005, 9 (1), 39.
(12) (a) Haver, A. C.; Smith, D. D. Tetrahedron Lett. 1993, 34, 2239. (b)
Boden, E. P.; Keck, G. E. J. Org. Chem. 1985, 50, 2394.
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1089

Table 2. Comparison of coupling reagents for the reaction of 2 and 3
HPLC area % 1
reagent
temp. (°C)
time (h)
solvent
equiv
HPLC area % 2
HPLC area % 3
ethyl ester
DIC
30
23
23
36
48
DMF CH2Cl2
DMF CH2Cl2
DMF
1.3
1.3
<1
<1
7
0
64
62
48
66
DCC
CDMT
BOP
30
trace
0-20
1.2 2 NMMa
1.15 2 Et3N
0
2
18
DMF
10
a NMM ) N-methylmorpholine.
Table 3. Comparison of 0.2 equiv of an auxiliary nucleophile
with 1.3 equiv of DIC in DMF at 4 h and 18-20 °C
HOBt was found in the isolated product when >0.2 equiv was
employed. Therefore, 0.2 equiv was used in all further scale-
up work.
auxiliary
nucleophile
HPLC
area % 2 area % 3
HPLC
HPLC area % 1
ethyl ester
Isolation of the ethyl ester or the corresponding ethyl ester
HCl salt of 1 could not be demonstrated in the laboratory. It
could be separated from the diisopropylurea byproduct and
HOBt via precipitation or extraction, but attempted isolation
gave rise to either an oil or a gummy solid with >10%
impurities. Chromatographic purification of the ethyl ester was
also studied, but again the results of this investigation were not
practical due to the insolubility of the product and competing
hydrolysis to the carboxylic acid. Conversion of the ethyl ester
to the corresponding sodium carboxylate and isolation of the
zwitterion species was therefore investigated.
Preparation of the sodium carboxylate analogue of 1 was
easily accomplished. Addition of 7.5 equiv of 2.5 N NaOH
directly to the coupling reaction completed the saponification
within 2 h. The product mixture was then extracted with
methylene chloride to remove some DMF, the isopropylurea
byproduct, and the HOBt before further manipulation.
Extensive experimentation was completed to define the
protocol for isolation of the zwitterion of 1. It was determined
that pH-controlled precipitation was required at or near the
zwitterion’s isoelectric point of 5.7. Laboratory studies found
that exceeding a pH of 7.5 or going below pH 3.8 caused a
large portion of the solids to dissolve, resulting in the product
falling out of the reaction mixture as a sticky, oily material.
However, when the pH was controlled within the range of
5.0-7.2, the desired zwitterion was precipitated as a white to
light-gray amorphous solid. Experimentally, this was ac-
complished by separate and simultaneous addition of the sodium
carboxylate product phase and 1.2 N HCl to a pH 7, aq 0.1 N
KH2PO4/NaOH buffer solution. A pH of approximately 7 was
maintained throughout the additions by controlling the rate at
which the streams were added. Once all of the carboxylate
solution was introduced, more 1.2 N HCl was added adjusting
the pH to 5.5-6.0. The desired zwitterion was collected by
filtration and dried, providing the product in 86-89% yield.
Two engineering concerns were recognized during the course
of this laboratory work. First, zwitterion prepared in this manner
coated the reactor walls, making isolation difficult. Fortunately
this issue was easily resolved by subsurface addition of both
the HCl and sodium carboxylate product streams. The second
concern observed was the physical properties and manipulation
of the wet cake. Cracks formed in the filter cake, preventing
the water level from decreasing to <50 wt %. Polyethylene bags
or rubber dams on the filter cakes helped slightly, but the drying
time to <15% LOD required days to complete. Confounding
this, the water would separate and the solids would turn into a
control: no auxiliary
2-hydroxypyridine
2-mercaptopyrdine
catechol
28
20
24
25
0.1
0.5
26
17
16
17
0.6
0.6
43
40
35
38
76
75
HOAt
HOBt
than in DMF, which in turn was greater than in 1:1 DMF/
CH2Cl2, but the combination of DMF/CH2Cl2 provided a more
efficient workup and isolation of the desired product. The ratios
of DMF/CH2Cl2 were also investigated, but again the 1:1
mixture provided the best balance with reaction results versus
postreaction isolation. DMAC was also investigated, but the
reaction rate was slower than in DMF or the DMF/CH2Cl2
combination.
The use of an auxiliary nucleophile8 accentuated the rate of
formation of the ethyl ester of 1 and gave rise to fewer
structurally related impurities. 1-Hydroxybenzotriazole (HOBt)
was found to be an excellent auxiliary nucleophile; however,
several chemical safety concerns were recognized.13 The
structurally related analogue, 1-hydroxy-7-azabenzotriazole
(HOAt), and other nucleophile auxiliaries were also studied,
but again there were questions about the safety of HOAt.
HOAt gave similar results when compared to HOBt (Table
3). The reaction was slightly faster, but the thermokinetic data
for both reagents revealed no safety advantage for HOAt. The
closed pan DSCs for HOBt and HOAt were similar.15 The other
nucleophiles, 2-hydroxypyridine, 2-mercaptopyridine, and cat-
echol, had slower reaction rates versus those of HOBt; thus, it
was decided to use the readily available and a safer form of
this reagent, a 12 wt % solution of HOBt in DMF,16 for further
scale-up work.
The addition of 0.1-0.4 mol equiv of HOBt17 to the DIC-
based coupling reaction accelerated the rate of reaction, decreas-
ing the reaction time from 23 to 4 h versus no catalytic HOBt.
It also gave rise to a higher reaction yield by decreasing the
number of impurities as measured by HPLC analysis. However,
(13) 1-Hydroxybenzotriazole: MSDS; Aldrich Chemical Company, 1999. For
a current MSDS see, for example: http://www.reagentworld.com/products/
msds2.asp?proid_2)18491.
(14) 1-Hydroxy-7-azabenzotriazole: MSDS; Aldrich Chemical Company,
1999. For a current MSDS see, for example: http://www.mctony.com/
hoat%20msds.htm.
(15) HOBt had an endotherm @ 155-158 °C and an exotherm @ 208 °C
(1772 J/g). HOAt had an exotherm at 212 °C (1789 J/g).
(16) Wehrstedt, K. D.; Wandrey, P. A.; Heitkamp, D. J. Hazard. Mater.
2007, A126, 1.
(17) HOBt is available as both a wet or dry reagent. It was found that up
to 36% water based on HOBt had no effect on the coupling reaction.
1090
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development

Table 4. Representative reaction or product mixture profile
discovered: anhydrous, a nonstoichiometric hydrate, and the
monohydrate 1. These pseudomorphic forms were found to arise
via a solution-mediated phase transformation with the desired
product 1 determined to be the thermodynamically most stable
form.19 Employing zwitterion obtained directly from the afore-
mentioned precipitation procedure, it was realized that impurities
did adversely effect the kinetics of various events during crystal-
lization, giving rise to 1 with purities ranging of 92.8 to 98.2
HPLC area %. However, with an API purity target set at a
minimum of 97.9% (93 wt %), it was decided that the zwitterion
would have to be chromatographed before conversion to 1.
To facilitate the technology transfer of the coupling reaction
and zwitterion precipitation process to the pilot plant as well
as ascertain the thermal and safety characteristics of the
chemistry, again reaction calorimetry experiments were com-
pleted.20 Table 5 lists a summary of the results. No concerns
were revealed. No gases or foaming were observed.
as measured by RP-HPLC analysis18
Pilot-Plant Preparation of 1. The aforementioned procedure
was completed on scale. In a typical batch, 35 kg of zwitterion
with a purity of 90% before chromatography was produced from
14.8 kg of 2 and 24.5 kg of 3. Drying of one batch was
attempted at 60 °C and 350 Torr using a 30 gallon glass-lined
tumble dryer, but the product adhered to the walls of the tumble
dryer and caramelized, contrary to the rotovap laboratory
studies. To avoid this, the wet cake was directly dissolved and
purified by column chromatography.
The zwitterion was purified on large scale. RP-HPLC
analysis was performed on each fraction to determine the area
% of product. Those fractions containing g96 area % were
combined and concentrated. After pH adjustment, the zwitterion
was collected by filtration, and the solids were dried.
Large-scale conversion of the zwitterion to 1 entailed mixing
the chromatographed, dried zwitterion with water and aceto-
nitrile. Without agitation, H3PO4 was added to the mixture and
heated to 60 °C. Heating without agitation was an important
operation, because mixing the slurry at ambient temperature
resulted in a cement-like material that literally stopped me-
chanical agitation! Heating without agitation afforded a flowable
mixture which quickly became a clear solution, making agitation
possible.
The antisolvent acetonitrile was added to the hot, stirred
solution over a 1.5 h period, producing a supersaturated solution.
However, nucleation did not occur during the charge, but rather
3-6 h after acetonitrile addition. Agitation and temperature were
maintained for 16-24 h to allow for the solution-mediated
phase transformation19 to the desired monohydrate polymorph
and to allow the system to attain its equilibrium concentration
at 60 °C. X-ray diffraction (PXRD) analysis confirmed the
formation of 1 (Figure 2). The slurry was cooled and the API
collected by filtration. The cake was washed and dried under
vacuum to provide 1 in 89-91% yield and 96.9 to 99.1 HPLC
area % purity.
glassy, caramelized mass. The annealed solids were very
difficult to work with, but HPLC analysis revealed no apparent
degradation. Nevertheless a procedure had to be developed that
would avoid these phenomena. In this regard, laboratory
pressure filtration and drying test were completed. After
extensive experimentation, it was found that the materials could
not be pressure filtered to dryness. A two step process had to
be followed. The zwitterion was first pressure filtered to a
45-48% moisture content, then immediately dried on a
rotoevaporator in Vacuo. In this manner an LOD of <4% could
be obtained providing gray, amorphous solids. This experience
suggested that the correct units of operation for drying on scale
would be pressure filtration followed by tumble drying. In
conjugation with the development of the methodology to isolate
the zwitterion of 1, the impurities were analyzed by HPLC
throughout the process in order to ascertain the effectiveness
of the procedure. A representative RP-HPLC profile of the
reaction mixture before and after precipitation of the zwitterion
as well as the final product 1 is provided in Table 4.
Most impurities were removed or diminished to <0.5 area
%, but none of the nonpolar impurities were significantly
decreased by the precipitation procedure. Thus, further API
purification would be required either before or during isolation.
Conversion of the zwitterion to 1 was extensively studied.19
Three predominate monophosphoric acid salt crystal forms were
The resulting product contained 2.4-2.8 wt % acetonitrile,
which could not be reduced further by drying! This level of
acetonitrile was a concern, since it was calculated to be too
close to ICH guidelines.21 Fortunately, it was found that the
(18) Clark, J. D. Unpublished results.
(19) (a) Nordstrom, F. L.; Rasmuson, A.; Sheikh, A. Y. J. Pharm. Sci.
2004, 93 (4), 995. (b) Sheikh, A. Y.; Schilke, T. C.; Czyzewski, A. M.;
Knable, C. M.; Hadley, A. J.; Danzer, G. D.; Iskos, V. H.; Moore,
C. M. V. Unpublished results.
(20) Reaction calorimetry was performed using a 2L A10 RC-1 reactor
equipped with a pitched blade impeller, ReactIR ATR FTIR probe, a
thermocouple, a calibration heater, and two addition funnels.
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1091

Table 5. Summary of the thermokinetic data for the DIC, catalytic HOBt peptide coupling of 2 and 3, ethyl ester
saponification, and zwitterion precipitation
total
heat (kJ)
intrinsic heat
(kJ/ mol x)
reaction
mass (kg)
heat capacity
(J/[kg K])
unit operation
ATR (°C)
dilution DMF and CH2Cl2
coupling reaction
saponification
-17.2
-18.3
-41.1
-30.0
-4.9, x ) DMF
-127, x ) 3
0.652
0.805
1.28
1560
1420
2680
4200
+16.9
+16.0
+12.0
+5.27
-38.0, x ) NaOH
-60.9, x ) HCl
precipitation
1.36
acetonitrile could be “exchanged” with acetone. Indeed, stirring
in excess acetone produced 1 in 97% yield containing 0.2-0.26%
acetonitrile and 3.1-3.4 wt % acetone.21 A representative
impurity profile of the final product 1 is listed in Table 4. A
photomicrograph of 1 is shown in Figure 3. Employing the
procedures detailed herein, 104 kg of GMP quality 1 was
prepared.
to-acetone exchange provided 104 kg of 1 as a crystalline
monohydrate with an average area % purity of 97.4% and in
67% overall yield from 4.
Experimental Section
(3S)-Glycyl-3-(3-bromo-5-chloro-2-hydroxyphenyl)-ꢀ-ala-
nine Ethyl Ester, Monohydrochloride (3). A reaction vessel
was charged with 4 (37.4 kg, 72.3 mol) and ethanol 2B (322
kg). The mixture was stirred and cooled to 0-5 °C. Acetyl
chloride (61.7 kg) was charged while maintaining the temper-
ature below 10 °C. The mixture was heated to 50 °C and held
for 3 h. The mixture was cooled to 20 °C and concentrated by
vacuum distillation. The concentrate was cooled to 35 °C, ethyl
acetate (300 kg) was added and the distillation repeated.
Additional ethyl acetate (250 kg) was charged and the distil-
lation repeated again. The concentrate was cooled to 25 °C and
ethyl acetate (337 kg) charged. The mixture was cooled to 5
°C and stirred for 3 h. The product was collected by filtration,
washed with precooled (5 °C) ethyl acetate (337 kg) and dried
at 50 °C until the LOD was measured at <3.0% providing 28
kg (93%) of 3. The product was shown to be identical to an
authentic sample by RP-HPLC, 13C NMR (DMSO-d6), and 1H
NMR (DMSO-d6) analyses.2
Conclusion
Studies directed toward the process research, development,
and scale-up preparation of the potential Rvꢀ3 integrin antagonist
1 were described. A procedure for the large-scale tert-butoxy-
carbonyl (BOC) deprotection of 4 producing the ꢀ-amino acid
ester 3 was provided. A convergent approach employing the
tetrahydropyrimidine hydroxybenzoic acid 2 linked to 3 via a
DIC, catalytic HOBt coupling reaction was detailed. Saponifica-
tion of the resulting ethyl ester, isolation of the corresponding
zwitterion, H3PO4 salt formation, crystallization, and acetonitrile-
(3S)-N-[3-Hydroxy-5-[(1,4,5,6-tetrahydro-5-hydroxy-2-
pyrimidinyl)amino]benzoyl]glycyl-3-(3-bromo-5-chloro-2-
hydroxyphenyl)-ꢀ-alanine. A reaction vessel was charged
with 2 (14.8 kg, 58.9 mol), 3 (24.5 kg, 58.9 mol), DMF
(89 kg) and 12 wt % 1-hydroxybenzotriazole (13 mol) in
DMF (15.0 kg). Methylene chloride (162 kg) was charged
to the mixture, which was stirred at 20 °C. N,N′-
Diisopropylcarbodiimide (9.6 kg, 76 mol) was charged
and the mixture stirred for 6 h at 20 °C. A second reaction
vessel was charged with water (153 L) and 50% sodium
hydroxide (35.3 kg). This solution was transferred to the
first reaction vessel and the mixture stirred for 2-2.5 h.
Methylene chloride (325 kg) was charged to the reaction
vessel and the mixture stirred. The phases were separated
and the aqueous phase washed twice with methylene
chloride (325 kg for each wash). In a separate vessel a
1.5 wt % aqueous KH2PO4 solution (162 kg water, 2.4 kg
KH2PO4) was prepared, and then the pH of the KH2PO4
solution was adjusted to between 6.5 and 7 with 2.5 N
NaOH (∼5.27 kg). To a third vessel was charged 4.3 wt
% HCl (141 L water, 19.8 kg of 37% HCl).
Figure 2. PXRD of the final product 1.
The aqueous phase in the first reaction vessel and 4.3 wt %
HCl solution were charged subsurface to the KH2PO4/NaOH
solution simultaneously such that the pH was maintained
between 6.0 and 7.2. When the addition was complete, the pH
of the resulting solution was further adjusted to 5.8 with the
4.3 wt % HCl solution. The product was collected by filtration,
washed with water (294 L), and dried on the filter with a warm
Figure 3. Photomicrograph of the final product 1.
1092
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development

stream of N2 to an LOD of 45-48% to give the corresponding
partially dried zwitterion (35 kg).
maintaining a temperature of 60 °C. The contents were
held at this temperature for at least 16 h. The crystalliza-
tion vessel was cooled to 5 °C over at least 5 h and the
product collected by filtration. The cake was washed with
acetonitrile (33.8 kg) and dried under vacuum until an
LOD assay of <4% was achieved.
A reaction vessel was charged with the zwitterion (35 kg)
and sufficient phosphoric acid, acetonitrile, and water to ensure
that there are two equivalents of phosphoric acid and a 50%
acetonitrile/water solution. The mixture was heated to 60 °C to
ensure dissolution and analyzed for the weight % of the
zwitterion using HPLC analysis. This value was used to
calculate the number of chromatography injections based on
the total amount of contained 1. Water was charged to achieve
a concentration of greater than 90% water as determined by
Karl-Fisher analysis. The mixture was purified on either a 2 in.
× 250 mm or a 4 in. × 1000 mm stainless steel column packed
with a YMC ODS-AQ, 50-µm silica reverse phase media using
phosphoric acid (H3PO4), acetonitrile, and water as eluting
solvents at a flow rate of 130 mL/min or 1000 mL/min,
respectively. After injecting the sample, the column was eluted
with 3.4 equivalent volumes of 90:10:0.2 water/acetonitrile/
H3PO4 (v:v:v). Those fractions containing e96 area % of 1
were combined and heated to 60 °C to ensure complete
dissolution of the product. The solution was filtered to remove
particulate matter and analyzed by HPLC to determine the wt
% of 1. The mixture was concentrated by vacuum distillation
to remove acetonitrile. The product was precipitated at 30 °C
using NaOH (7.6 wt %) until a pH of 5.5 ( 0.5 units was
obtained. The mixture was cooled to 25 °C and allowed to stir
for 2-3 h. The product was collected by filtration, washed with
water, and dried on the filter using a warm stream of N2 for
21 h at 350 Torr until the water content was e5.0% as
determined by KF analysis to give 22.6 kg (62.4%) of the
zwitterion. The product was shown to be identical to an
authentic sample by RP-HPLC, 13C NMR (DMSO-d6), and 1H
NMR (DMSO-d6) analyses.2
The crude 1 (21.5 kg) was charged to a reaction vessel.
Acetone (153 kg) and water (21.5 kg) were added, and
agitation was initiated. The contents of the vessel were
heated to 40 °C and held for 3 h. The mixture was cooled
to 20 °C and the product collected by filtration and dried
on the filter with a warm stream of N2 for 31 h at 350
Torr to give 21.0 kg (84%) of 1 containing <0.26 wt %
acetonitrile and 3.6 wt % acetone as determined by GC
analysis. The water content was measured as 3% as
determined by KF analysis. The product was shown to be
identical to an authentic sample by RP-HPLC, 13C NMR
1
(DMSO-d6), H NMR (DMSO-d6), and PXRD analyses.2
Acknowledgment
We thank the many operators and support personnel of the
former Pharmacia Queeny and Skokie pilot plants. Also
acknowledged are Teddy Albano, Lois Amari, Leslie Askonas,
Jessica Ballinger, Kathy Barton, Jim Behling, Kerry Brown,
Steve Brown, Joe Collins, A. J. Damascus, Gerald Danzer, Steve
Doherty, John Dygos, Ray Frame, Nancy Gallagher, Mary
Gasper, A. J. Hadley, D. Herren, Nick Hoffmeister, Erwin
Irdam, Wei Kong, Frank Kumli, Chin Liu, Shan Lin, Jim Miles,
Larry Miller, Thomas Muller, John Ng, Sean Nugent, Vlad
Orlovski, Greg Pertle, Dan Pilipauskas, Joost Quaadgras, Ed
Sheldon, Bruce Sorensen, Glenn Stahl, Don Stano, L. J.
Streicher, Kara Teklinski, Joe Timko, Steve Trank, Zhi Wang,
the Chemical Sciences Analytical Group, the Chemical Sciences
Preparative Chromatography Group, and the Pharmaceutical
Analytical Sciences Group.
(3S)-N-[3-Hydroxy-5-[(1,4,5,6-tetrahydro-5-hydroxy-2-
pyrimidinyl)amino]benzoyl]glycyl-3-(3-bromo-5-chloro-2-
hydroxyphenyl)-ꢀ-alanine Monophosphate (1:1) (Salt) Mono-
hydrate (1). A crystallization vessel was charged with
water (86.0 kg) and the zwitterion (21.5 kg). Without
agitation, a solution of acetonitrile (67.6 kg) and phos-
phoric acid (4.25 kg) was charged. The mixture was heated
to 65 °C. Once all the solids dissolved, agitation was
started and additional acetonitrile (135 kg) charged while
Received for review July 9, 2009.
OP900173F
(21) http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatory-
Information/Guidances/ucm073395.pdf.
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1093