Process for the preparation of an amorphous, peptide-like diabetes drug: Approach to a chromatography-free process

Article abstract of DOI:10.1021/op100084r

A manufacturing process suitable for large-scale production of the peptide-like amorphous compound N-((2R)-1-{(3R)-6-chloro-3-[(dimethylamino) methyl]-3,4-dihydroquinolin-1(2H)-yl}-3-(1H-indol-3-yl)-1-oxopropan-2-yl) -1-[(1-methyl-1H-indol-2-yl)carbonyl]p

Full text of DOI:10.1021/op100084r

Organic Process Research & Development 2010, 14, 1110–1117
Process for the Preparation of an Amorphous, Peptide-like Diabetes Drug: Approach
to a Chromatography-Free Process
Yasuhiro Sawai,* Taihei Yamane,† Motoki Ikeuchi, Shinji Kawaguchi, Masatoshi Yamada, and Mitsuhisa Yamano
Chemical DeVelopment Laboratories, CMC Center, Takeda Pharmaceutical Company Limited, 17-85, Jusohonmachi 2-chome,
Yodogawa-ku, Osaka 532-8686, Japan
Abstract:
but to develop them as amorphous active pharmaceutical
ingredients (APIs).7 In this case, the purification method of
amorphous APIs has become a pivotal issue for their process
development because conventional purification such as that
using silica gel chromatography is undesirable for large-scale
production.8 Alternative purification methods suitable for large-
scale manufacturing may include extraction-based purification
and certain types of chromatography with a recyclable stationary
phase, such as ion-exchange resin, chelate resin, and synthetic
absorbent. In most cases, a chromatography-free process is
considered the most desirable, although it is difficult to develop.9
In the course of a drug discovery study for diabetes mellitus,
N-((2R)-1-{(3R)-6-chloro-3-[(dimethylamino)methyl]-3,4-dihy-
droquinolin-1(2H)-yl}-3-(1H-indol-3-yl)-1-oxopropan-2-yl)-1-
[(1-methyl-1H-indol-2-yl)carbonyl]piperidine-4-carboxamide
(1), which has a tetrapeptide-like structure with three peptide
bonds, was discovered by our discovery team.10 API 1 was
obtained only as a stable amorphous solid that has relatively
high glass-transition temperature (103-105 °C). The discovery
synthesis of 1 started from the preparation of chiral tetrahyd-
roquioline fragment (S)-2 Via an enzymatic esterification process
(Scheme 1).10 The resultant (S)-2 was converted to syn-3a Via
peptide coupling with Fmoc-D-tryptophan followed by Fmoc-
deprotection. The subsequent peptide coupling of syn-3a with
carboxylic acid 4 provided 1 in gram quantities. Amorphous
API 1 was purified by silica-gel chromatography and alumina
chromatography, which are to be avoided in large-scale produc-
tion if possible.
A manufacturing process suitable for large-scale production of
the peptide-like amorphous compound N-((2R)-1-{(3R)-6-chloro-
3-[(dimethylamino)methyl]-3,4-dihydroquinolin-1(2H)-yl}-3-(1H-
indol-3-yl)-1-oxopropan-2-yl)-1-[(1-methyl-1H-indol-2-yl)carbon-
yl]piperidine-4-carboxamide (1) as a drug for treating diabetes has
been developed. The first kilogram quantities of 1 were prepared
Wia a single-chromatography process that employed recyclable
cation-exchange resin chromatography for an amorphous inter-
mediate (2R)-2-amino-1-[(3R)-6-chloro-3-[(dimethylamino)methyl]-
3,4-dihydroquinolin-1(2H)-yl]-3-(1H-indol-3-yl)-propan-1-one (syn-
3a). We have also developed a chromatography-free process that
involves a combination purification of extraction of syn-3a with
crystallization of syn-3a ·0.25H3PO4 ·0.5H2O. The latter process
afforded amorphous compound 1 with >98% purity by
HPLC area analysis, the same quality as that provided by
the former process.
Introduction
Peptide-based drug molecules are prevalent in drug discovery
studies that target receptors or enzymes.1-3 In most cases, a
lead peptide is converted into a low-molecular weight peptide
mimetic compound with peptide-like structure by reducing the
number of peptide bonds and introducing general organic
molecules other than natural amino acids.4,5 The peptide-like
compounds can acquire more favorable pharmaceutical proper-
ties than those of the original molecules, such as high stability
against proteolytic degradation, good absorption in the gas-
trointestinal tract after oral ingestion, and slow excretion through
liver and kidneys.6 However, they often exhibit unfavorable
physical properties; for example, they are frequently obtained
as amorphous solids with poor solid properties.
In order to support preclinical pharmacological and toxico-
logical evaluations and potential clinical applications, it was
necessary to develop a manufacturing process for multikilogram
quantities of 1. The most critical issue in process development
was the purification of amorphous API 1. In addition to this
challenge, the purification of the penultimate intermediate, (2R)-
When peptide like compounds and their salts are not
available in crystalline or cocrystalline form, there is no choice
(7) For the development of crystallizable compounds as amorphous APIs
as a means of improving their pharmaceutical delivery, see: Lian, Y.
AdV. Drug DeliVery ReV. 2001, 48, 27–42.
* Author to whom correspondence may be sent. E-mail: Sawai_Yasuhiro@
takeda.co.jp.
(8) For large-scale preparative chromatography, see: (a) Welch, C. J.;
Fleits, F.; Antia, F.; Yehl, P.; Waters, R.; Ikemoto, N.; Armstrong,
J. D.; Mathre, D. J. Org. Process Res. DeV. 2004, 8, 186–191. (b)
Welch, C. J.; Henderson, D. W.; Tschaen, D. M.; Miller, R. A. Org.
Process Res. DeV. 2009, 13, 621–624. (c) Schmidt-Traub, H., Ed.
PreparatiVe Chromatography of Fine Chemicals and Pharmaceutical
Agents; Wiley-VCH: Weinheim, 2005.
† Current address: Production Control Department, Pharmaceutical Production
Division, Takeda Pharmaceutical Company Limited, 1-1, Doshomachi 4-chome,
Chuo-ku, Osaka 540-8645, Japan.
(1) Weckbecker, G.; Lewis, I.; Albert, R.; Schmid, H. A.; Hoyer, D.;
Bruns, C. Nat. ReV. Drug DiscoVery 2003, 2, 999–1017.
(2) Blakeney, J. S.; Reid, R. C.; Le, G. T.; Fairlie, D. P. Chem. ReV. 2007,
107, 2960–3041.
(9) For an efficient chromatography-free process of tetrapeptide, thyrotro-
pin-releasing hormone, see: Hatanaka, C.; Obayashi, M.; Nishimura,
O.; Toukai, N.; Fujino, M. Biochem. Biophys. Res. Commun. 1974,
60, 1345–1350.
(3) Tyndall, J. D. A.; Pfeiffer, B.; Abbenante, G.; Fairlie, D. P. Chem.
ReV. 2005, 105, 793–826.
(4) Wu, Y.-D.; Gellman., S. Acc. Chem. Res. 2008, 41, 1231–1232.
(5) Gante, J. Angew. Chem., Int. Ed. Engl. 1994, 33, 1699–1720.
(6) Giannis, A.; Kolter, T. Angew. Chem., Int. Ed. Engl. 1993, 32, 1244–
1267.
(10) (a) Abe, H. MEDCHEM NEWS 2007, 17 (3), 18–22. (b) Kato, K.;
Terauchi, J.; Suzuki, N.; Takekawa, S. PCT Int. Appl. WO/2001/
025228(A1), 2001.
1110
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development
10.1021/op100084r  2010 American Chemical Society
Published on Web 07/23/2010

Scheme 1. Discovery synthesis of 1
Scheme 2. Synthesis of (S)-2
generation) process could provide 1 of the same quality as that
of the one provided by the former process without any
chromatography. Herein, we describe the process development
of peptide-like amorphous API 1 with particular focus on the
purification of amorphous intermediate syn-3a.11
2-amino-1-[(3R)-6-chloro-3-[(dimethylamino)methyl]-3,4-dihy-
droquinolin-1(2H)-yl]-3-(1H-indol-3-yl)-propan-1-one (syn-3a),
was a major issue, because it was also obtained as an amorphous
solid. If high-quality syn-3a was not prepared, a significant
amount of impurities derived from syn-3a would remain in 1.
These impurities would have a profound impact on the devel-
opment of 1.
Our preliminary study showed that the coupling reaction of
syn-3a with 4 proceeded relatively cleanly to produce 1 and
that syn-3a was obtained with a considerable amount of
byproduct, which included a large amount of the residue derived
from the Fmoc protecting group. On the basis of these results,
we expected that the chromatography at the final step could be
omitted if the final coupling reaction was carried out with the
highest possible purity syn-3a under appropriate conditions. In
contrast, it seemed difficult to prepare high-quality syn-3a
without chromatographic purification. From the viewpoint of
large-scale manufacturing, even if there were no other alternative
purification methods except chromatography for the amorphous
compounds, it is desirable that the recyclable resin chromatog-
raphy of the most efficient separation mode be incorporated
into the process at the most effective step.
Hence, we initially approached a single-chromatography
process by formulating a purification strategy for the entire
process as follows: (1) syn-3a was purified at as high a level as
possible by recyclable resin chromatography, (2) the generation
of impurities at the final step was minimized by optimizing the
reaction conditions, and (3) API 1 was purified by extraction
without chromatography. As a further study, we approached a
chromatography-free process by replacing the chromatography
with extraction-based purification of syn-3a. The former (first
generation) process successfully produced the kilogram quanti-
ties of 1 necessary for preclinical evaluations. The latter (second
Results and Discussion
Synthesis of (S)-2. The discovery synthesis Via the enzy-
matic esterification process required 13 steps for the preparation
of (S)-2 (Scheme 1). In order to develop a shorter process, we
attempted optical resolution Via diastereomeric salt formation
of rac-2, which can be prepared in five steps (Scheme 2).
Knoevenagel condensation of commercially available alde-
hyde 5 with diethyl malonate in the presence of sodium acetate
in acetic anhydride quantitatively provided R,ꢀ-unsaturated
diester 6.12,13 After typical extraction workup, crude oil 6 was
used without a further purification in the next step. The reduction
of both double bond and esters with sodium tetrahydroborate
afforded saturated diol 7 in 77% yield.14 After bismethane-
sulfonylation of diol, the nitro group of 8 was reduced using
zinc in acetic acid. This reaction proceeded chemoselectively
and was accompanied with in situ intramolecular cyclization.
The resultant cyclic intermediate reacted with dimethylamine
to give rac-2 in 77% yield.
Then, various types of chiral acids were evaluated for their
potential utility as resolving agents for rac-2. The screening
revealed that N-Ts-leucine 9 was the most effective for resolving
rac-2. Interestingly, the configuration of preferentially crystal-
lized salt depended on the solvent used. For example, (R)-9
(11) Sawai, Y. Process Research of Peptide Mimetic Compounds as
Diabetic Drug Candidate, Proceedings of The Eighteenth International
Conference on Organic Process Research and DeVelopment, The
Eighteenth International Conference on Organic Process Research and
Development, Montreal, Canada, June 23-26, 2008.
(12) Suzuki, M.; Ohuchi, Y.; Asanuma, H.; Kaneko, T.; Yokomori, S.;
Ito, C.; Isobe, Y.; Muramatsu, M. Chem. Pharm. Bull. 2000, 48, 2003–
2008.
(13) Lewandowska, E.; Kinastowski, S.; Wnuk, S. F. Can. J. Chem. 2002,
80, 192–199.
(14) Curran, D.; Gabarda, A. E. Tetrahedron 1999, 55, 3327–3336.
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1111

Scheme 3. Resolution of rac-2 Wia diastereomeric salt
formation with (R)-9 or (S)-9
Scheme 4. Synthesis of syn-3a
gave the salt with (S)-2 from ethyl acetate and the salt with
(R)-2 from ethanol. Proton NMR study demonstrated that using
(R)-9 in ethyl acetate gave (S)-2·(R)-9 as a solvent-free crystal,
while using (S)-9 in ethanol gave (S)-2·(S)-9 as a solvate of
ethanol (Scheme 3).15 Under the optimum conditions, (S)-
2·(R)-9 or (S)-4·(S)-9·0.2EtOH was obtained in 31% yield with
89% de or in 31% yield with 73% de, respectively. The
resolution efficiency with (R)-9 was higher than that with (S)-
9. The subsequent salt splitting of (S)-2·(R)-9 followed by the
crystallization of free base from 1:2 toluene/n-hexane improved
the optical purity to give (S)-2 in 85% yield with 99% ee.16 It
should be noted that the synthetic process of 1 did not involve
any crystallization purification after the introduction of chiral
parts, and thus the undesired epimer was difficult to remove.17
The preparation of high optical purity (S)-2 was one of the key-
points for the successful preparation of high-quality 1. Therefore,
we carried out additional recrystallization of (S)-2 that achieved
further improvement of the optical purity to >99.9% ee.
This process employed potentially hazardous nitro com-
pounds; therefore, we conducted a safety evaluation before the
first scale-up campaign. Differential scanning calorimetry (DSC)
analysis of four nitro compounds revealed that they did not
produce significant safety issues arising from their thermo-
chemical behavior.18 Therefore, we carried out pilot-plant
Table 1. Optimization of the coupling reaction of (S)-2 with
(R)-10a
base
entry (R)-10 (equiv)
equiv syn-3b yield (%)
1
2
3
4
5
3.0
3.0
3.0
1.5
1.5
NaOHb
i-Pr2NEt
i-Pr2NEt
i-Pr2NEt
Et3N
1.1
1.1
1.6
1.4
1.4
87
57
86
87
90
a 1.2 equiv of (COCl)2 was used. b 0.1 equiv of n-Bu4NHSO4 was added.
campaigns, and multikilogram quantities of (S)-2 were suc-
cessfully obtained through the seven-step process.19
Synthesis of syn-3a. The next task was to improve the
coupling reaction of (S)-2 with N-Fmoc-D-tryptophan (R)-10
(Scheme 4). Improvement of the reaction can reduce the load
on the purification of amorphous compound syn-3a. This
reaction required strong activation for the carboxyl function of
D-tryptophan, such as that provided by acyl chloride, because
the ring nitrogen of (S)-2 has a low nucleophilicity arising from
the 6-chloro substituent on the tetrahydroquinoline ring.20 The
Fmoc protective group was selected for the R-nitrogen protec-
tion of D-tryptophan because of its stability under acidic
conditions, which was derived from acyl chloride and its
removability under mild basic conditions. Although syn-3b was
obtained in 87% yield under the discovery synthesis conditions
(Table 1, entry 1), poor reproducibility was observed. The use
of strongly basic sodium hydroxide was suspected to promote
the hydrolysis of acyl chloride intermediate thus leading to the
inconsistent results. Therefore, we attempted to use less basic
(15) Powder X-ray diffraction of each of (S)-2·(R)-9 and (R)-2·(R)-9, which
were prepared from enantiopure (S)-2 and (R)-2 in ethanol or ethyl
acetate, demonstrated that (S)-2·(R)-9 consistently showed the same
pattern, while (R)-2·(R)-9 showed two types of patterns, depending
on the solvent used. For the resolution of both enantiomers Via
diastereomeric salt formation using a single configuration of resolving
agent, see: (a) Hirose, T.; Begum, M.; Islam, M. S.; Taniguchi, K.;
Yasutake, M. Tetrahedron: Asymmmetry 2008, 19, 1641–1646. (b)
Taniguchi, K.; Aruga, M.; Yasutake, M.; Hirose, T. Org. Biomol.
Chem. 2008, 6, 458–463. (c) Taniguchi, K.; Sakurai, R.; Sakai, K.;
Yasutake, M.; Hirose, T. Bull. Chem. Soc. Jpn. 2006, 79, 1084–1090.
(d) Sakai, K.; Sakurai, R.; Nohira, H.; Tanaka, R.; Hirayama, N.
Tetrahedron: Asymmmetry 2004, 15, 3495–3500. (e) Sakai, K.;
Sakurai, R.; Hirayama, N. Tetrahedron: Asymmmetry 2004, 15, 1073–
1076.
(16) The enantiomeric excess of 2 was determined by HPLC. HPLC
conditions: column, Chiralpak AD-RH (4.6 mm × 150 mm); mobile
phase, 0.05 M aq KH2PO4 containing 0.1% of Et3N (pH 7.5)/MeCN
(50:50); flow rate, 0.5 mL/min; column temperature, 25 °C; detection,
UV 254 nm.The recrystallization of (S)-2·(R)-9 of 88% de in ethyl
acetate gave the product in 82% yield with 99% de.
(17) Use of (S)-2 of low optical purity led to the formation of syn-3a in
low diastereomeric ratio, because anti-3b was generated much faster
than syn-3b and the undesired anti-isomer was difficult to remove
from the desired syn-isomer.
(18) DSC analysis shows that these nitro compounds have sufficiently high
exothermic onset temperatures (5, 218 °C; 6, 328 °C; 7, 302 °C; 8,
252 °C), although they have large gross heating values (5, 1513 J/g;
6, 363 J/g; 7, 755 J/g; 8, 1091 J/g). Comprehensive safety evaluations
should be conducted before further attempts to scale up.
(19) It should be noted that reduction reaction using sodium borohydride
in alcohol solvent can cause safety concerns including latent and abrupt
release of heat. For the control of heat generation on large-scale usage
of sodium borohydride in alcohol, see: (a) Flanagan, R. C.; Xie, S.;
Millar, A. Org. Process Res. DeV. 2008, 12, 1307–1312. (b) Belecki,
K.; Berliner, M.; Bibart, R. T.; Meltz, C.; Ng, K.; Phillips, J.; Ripin,
D. H. B.; Vetelino, M. Org. Process Res. DeV. 2007, 11, 754–761.
(20) The reactions of (S)-2 with N-protected D-tryptophans (N-Fmoc-D-Trp-
OH, N-Boc-D-Trp-OH or N-Cbz-D-Trp-OH) using condensation reagent
1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride (EDC·HCl)
did not give the desired amides. For the intramolecular N-acylation of
tetrahydroquinoline using N,N′-dicyclohexylcarbodiimide (DCC), see:
Somayaji, V.; Brown, R. S. J. Org. Chem. 1986, 51, 2676–2686.
1112
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

tertiary amine instead of sodium hydroxide. Using 1.1 equiv
of diisopropylethylamine (DIEA) with 1.2 equiv of oxalyl
chloride, however, caused the yield of syn-3b to be significantly
decreased (entry 2). After extensive investigation into the
amount of reagents, we found that using approximately 1.5
equiv of DIEA with 1.2 equiv of oxalyl chloride brought the
yield back to the former level (entry 3). The acyl chloride
formation and the subsequent coupling reaction could be
conducted in one-pot. Although a large excess of (R)-10 was
necessary for completion of the coupling reaction under the
discovery synthesis conditions (entry 1), the amount of (R)-10
was successfully reduced to 1.5 equiv without a yield decrease
(entry 4), and even in the case using triethylamine syn-3b was
obtained in 90% yield (entry 5). Applying the optimum
conditions, an efficient and robust process was accomplished.
After typical extraction workup, syn-3b was used without a
further purification in the subsequent Fmoc-deprotection reaction
(Scheme 4). Although the Fmoc protective group worked
effectively during the coupling reaction,21 dibenzofulvene 11
and its piperidine adduct 12 appeared during deprotection and
were difficult to remove. Success in reducing of the amount of
(R)-10 at the coupling reaction step was also advantageous for
the purification of amorphous compound syn-3a, because the
amounts of 11 and 12 were consequently reduced. However,
syn-3a still contained more than a stoichiometric amount of
these byproduct, almost a half molar amount of N-H-D-
tryptophan derived from the excess (R)-10, and a wide variety
of low-level impurities.
Cation-Exchange Resin Chromatography of syn-3a. Ac-
cording to the purification strategy, we approached recyclable
resin chromatography for the removal of these impurities. The
screening of the separation mode of the chromatography using
cation-exchange resin, chelate resin, and synthetic adsorbent
revealed that cation-exchange resin chromatography (CEC) was
effective for the purification of syn-3a. In particular, the weakly
acidic cation-exchange resin DIAION WK100, which consists
of a methacrylate resin matrix, was the most efficient stationary
phase. Figure 1 illustrates the concept behind the CEC procedure
for syn-3a: (1) a crude mixture was charged onto the column,
which had been filled with DIAION WK100 preconditioned
in its H-form,22 (2) impurities were eluted while syn-3a
remained adsorbed to the H-form resin, and (3) syn-3a was
exchanged with sodium cations and eluted from Na-form resin.
The most significant issue for the development of the CEC
procedure was the separation of tertiary amine syn-3a and 12.
An extensive investigation into the concentration of sodium
cations in mobile phase revealed that the sodium cations in 0.5%
sodium chloride aqueous solution/methanol (20:80) selectively
exchanged 12, while the sodium cations in 5% sodium chloride
aqueous solution/methanol (20:80) simultaneously exchanged
both syn-3a and 12. The CEC procedure has been developed
Figure 1. CEC for syn-3a
on the basis of this selective cation exchange. After eluting the
neutral substances such as 11 and N-H-D-tryptophan by
methanol, 5% sodium chloride aqueous solution/methanol (20:
80) eluted 12 with low levels of other impurities. Then, 5 M
sodium hydroxide aqueous solution/methanol (20:80) eluted
syn-3a. This fraction gave amorphous solid syn-3a with 97%
purity by HPLC area analysis and 73% yield based on (S)-2.
This process has been run on a kilogram scale, and the resin
could be reused at least three times after reconditioning, in which
the resin was converted from its Na-form to its H-form using
hydrochloric acid and methanol. As discussed in more detail
below, syn-3a was converted to 1 with >98% HPLC area.
Therefore, we have developed a first-generation process of 1
suitable for large-scale production, which involved single
chromatography throughout the entire process.
Extraction-Based Purification of syn-3a. Before the CEC
was developed, all attempts to crystallize the salt of syn-3a with
various types of acid were unsuccessful. However, when
applying the high-quality amorphous solid syn-3a, which was
purified by the CEC, to the salt crystallization study, the
phosphate salt was successfully isolated as a stable crystal. The
composition of the phosphate salt was determined by single-
crystal X-ray analysis (Figure 2, top and side views of the
phosphate salt). It was revealed that the crystal consists of four
molecules of syn-3a, a molecule of phosphoric acid, and two
molecules of water. One of the two molecules of water
coordinates the phosphoric acid. The resulting four of the five
hydrogen atoms interact with the nitrogen atoms of four
molecules of syn-3a, and one of the five hydrogen atom interacts
with an another water. Using 4 equiv of phosphoric acid in
aqueous ethanol on the basis of the composition of the
phosphate salt, crystalline syn-3a ·0.25H3PO4 ·0.5H2O was
obtained in the highest yield. The combination of the CEC with
the subsequent phosphate salt crystallization enabled preparation
of syn-3a ·0.25H3PO4 ·0.5H2O with >99% HPLC area in 64%
yield based on (S)-2.
(21) For the peptide synthesis Via amino acid chloride using Fmoc protective
group, see: (a) Carpino, L. A.; Beyermann, M.; Wenschuh, H.; Bienert,
M. Acc. Chem. Res. 1996, 29, 268–274. (b) Carpino, L. A. Acc. Chem.
Res. 1987, 20, 401–407.
(22) The weakly acidic cation-exchange resin DIAION WK100 was pre-
conditioned as follows: A column was filled with DIAION WK100
(100 L) using methanol and was successively eluted with the mobile
phase as follows: water/methanol (200 L, 1:4), water/methanol (200
L, 1:1), water/methanol (200 L, 4:1), water (200 L), 1 M hydrochloric
acid (300 L), water (300 L), and methanol (300 L).
Success in the salt crystallization allowed the preparation
of highly purified syn-3a without chromatography. One of the
possible methods was the use of extraction purification in
combination with the phosphate salt crystallization. We inves-
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1113

Figure 2. Single-crystal X-ray analysis of syn-3a ·0.25H3PO4 ·0.5H2O.
Scheme 5. Synthesis of 1
purification step, we minimized the impurities generated by
optimizing the coupling reaction conditions. The coupling
reactions using oxalyl chloride, pivaloyl chloride, ethyl chlo-
roformate, or N, N′-carbonyldiimidazole (CDI) resulted in the
formation of 1 with a considerable amount of byproduct. In
contrast, with the use of condensation reagent 1-[3-(dimethy-
lamino)propyl]-3-ethylcarbodiimide hydrochloride (EDC ·HCl)
with 1-hydroxybenzotriazole monohydrate (HOBt), the reaction
was completed in a short time to give 1 with improved purity.
The best result was obtained with 1.2 equiv each of 4,
EDC ·HCl, and HOBt in DMF at ambient temperature. In
particular, using 1.2 equiv of 4 was critical for the complete
consumption of syn-3a, which was difficult to remove by
extraction.
After the final peptide coupling reaction was carried out
under the optimum conditions, extraction using aqueous potas-
sium carbonate solution and ethyl acetate afforded amorphous
powder 1 with >98% HPLC area from syn-3a or syn-3a
·0.25H3PO4 ·0.5H2O without chromatographic purification. The
residual 4 was removed by washing of organic layer with 5%
aqueous potassium carbonate solution. After decolorization of
the organic layer with activated carbon, solvent was switched
to ethanol by vacuum concentration with additional ethanol. A
simple and scalable precipitation was also developed: dropping
10 volumes of ethanol solution of 1 into 50 volumes of water
at ambient temperature effectively provided amorphous powder.
The wet powder of 1 was dried under reduced pressure until
the water content is not more than 2% by Karl Fisher method.
Figure 3. Extraction of syn-3a.
tigated the extraction conditions using a weakly acidic aqueous
phase on the basis of the concept of CEC. Adjusting the aqueous
phase to pH 6, syn-3a was effectively separated from major
impurities 11 and 12 (Figure 3). Then, syn-3a was extracted
with ethyl acetate after adjusting the aqueous phase to pH 11.
N-H-D-Tryptophan was removed by a sequence of the extrac-
tion. Unfortunately, a wide variety of low-level impurities were
carried downstream, and therefore, amorphous compound syn-
3a was obtained with a purity of only 88% by HPLC area
analysis. However, the subsequent phosphate salt crystallization
from aqueous ethanol followed by an additional purification
by reslurrying in 3:1 ethyl acetate/n-hexane successfully im-
proved the purity to give syn-3a ·0.25H3PO4 ·0.5H2O with
>99% HPLC area in 68% yield based on (S)-2. Therefore, a
second-generation chromatography-free process has been
developed.
Synthesis of 1. The synthesis of 1 was completed with a
peptide-coupling reaction of amine syn-3a with carboxylic acid
4 (Scheme 5), which was readily prepared with high purity from
commercially available ethyl isonipecotinate with 1-methyin-
dolyl-2-carboxylic acid. In order to reduce the burden on the
1114
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

Applying the optimum conditions, the pilot-plant campaigns
successfully provided the kilogram quantities of amorphous API
1 in 87% yield with >98% HPLC area and >99.9% ee.23 The
alternative extraction-based purification of syn-3a allowed 1 to
be prepared, without any chromatography, and of a quality
matching that provided by the process involving CEC.
kg). The combined organic layers were washed with water (3
× 477.6 kg), saturated aq NaHCO3 (2 × 425.9 kg), and 5% aq
NaCl (2 × 417.9 kg), and then concentrated in Vacuo. Ethanol
(158.0 kg) was added to the residue, and the resultant solution
was concentrated in Vacuo (twice). Then ethanol (126.6 kg)
was added to the residue to give a clear solution. Under nitrogen
atmosphere, to another reaction vessel containing ethanol (379.3
kg) was added sodium tetrahydroborate (24.4 kg, 645.0 mol)
in small portions at -10-0 °C. The solution obtained above
was added dropwise to the reaction vessel containing ethanol
and sodium tetrahydroborate. The mixture was allowed to warm
and was stirred at room temperature for 2.5 h. After cooling to
0 °C, 6 M hydrochloric acid (206.6 kg) was added dropwise.
The mixture was allowed to warm to room temperature, and
water (238.8 kg) was added. After stirring at room temperature
for 1.5 h, the mixture was adjusted to pH 6 by adding 30% aq
NaOH (48.1 kg) and then was concentrated to 320 L in Vacuo.
Water (238.8 kg) and tert-butylmethylether (235.6 kg) were
added, and the layers were separated. The aqueous layer was
extracted with tert-butylmethylether (235.6 kg). The combined
organic layers were washed with water (2 × 398.0 kg) and
then concentrated in Vacuo. Toluene (68.9 kg) was added to
the residue, and the resultant mixture was concentrated in Vacuo
(twice). After the addition of toluene (79.2 kg), the mixture was
heated to 40 °C and stirred until a clear solution was obtained.
Then the mixture was stirred at room temperature for 2 h and
at 0-5 °C for 1.5 h. The resultant precipitate was collected by
filtration, washed with cold toluene (34.6 kg), and dried in Vacuo
to give 7 (40.4 kg, 164.5 mol, 77% yield) as a white crystalline
powder. Mp 59-60 °C; IR (ATR) ν 3282, 1519, 1333, 1028,
831 cm-1; MS (ESI) m/z 246 (M + H)+; 1H NMR (300 MHz,
CDCl3) δ 2.02-2.09 (m, 1H), 2.43 (t, J ) 5.0 Hz, 2H), 3.00
(d, J ) 7.3 Hz, 2H), 3.67-3.74 (m, 2H), 3.83-3.89 (m, 2H),
7.35 (dd, J ) 2.3, 8.7 Hz, 1H), 7.43 (d, J ) 2.3 Hz, 1H), 7.92
(d, J ) 8.7 Hz, 1H); 13C NMR (75 MHz, CDCl3) δ 31.0, 42.9,
64.3, 126.6, 127.7, 132.9, 137.6, 139.3, 146.8; Anal. Calcd for
C10H12NO4Cl: C, 48.89; H, 4.92; N, 5.70; Cl, 14.43. Found: C,
48.92; H, 4.86; N, 5.76; Cl, 14.36.
Synthesis of 2-(5-Chloro-2-nitrobenzyl)propane-1,3-diyl
Dimethanesulfonate (8). To a solution of 7 (40.3 kg, 164.0
mol) in ethyl acetate (181.8 kg) were successively added
triethylamine (49.8 kg, 492.1 mol) and methanesulfonyl chloride
(56.4 kg, 492.4 mol) at 0-10 °C. The mixture was stirred at
0-10 °C for 1 h and then allowed to warm to room temperature.
After adding saturated aq NaHCO3 (215.5 kg), the mixture was
stirred at room temperature for 1 h and then extracted with ethyl
acetate (181.8 kg). The organic layer was washed with saturated
aq NaHCO3 (215.5 kg), 10% aq NaCl (221.7 kg), 2 M
hydrochloric acid (215.5 kg), 10% aq NaCl (221.7 kg), and
5% aq NaCl (6 × 211.8 kg) and then was concentrated in
Vacuo. Ethyl acetate (109.0 kg) was added to the residue, and
the resultant mixture was allowed to cool to 0 °C and stirred
for 1 h. Then diisopropylether (87.6 kg) was added dropwise,
and the mixture was stirred at 0 °C for 2 h. The resultant
precipitate was collected by filtration, washed with 1:1 ethyl
acetate/diisopropylether (65.4 kg), and dried in Vacuo to give
8 (58.8 kg, 146.3 mol, 89% yield) as a white crystalline powder.
Mp 73-74 °C; IR (ATR) ν 1522, 1339, 1173, 945, 823, 522
Conclusion
The process of manufacturing peptide-like API 1 suitable
for large-scale production is required to support preclinical
evaluations and potential clinical applications. The most critical
issue to be addressed for the process development was the solid-
state properties of 1 and the final intermediate syn-3a. These
are both amorphous solids that are difficult to purify in scaled-
up production. We initially considered a single-chromatography
process that employed recyclable cation-exchange resin chro-
matography for syn-3a. The process successfully provided
kilogram quantities of 1 with >98% HPLC area. As a further
study, we attempted to replace the CEC with extraction-based
purification. The chromatography-free process that involved the
extraction purification of syn-3a and the successive crystalliza-
tion of syn-3a ·0.25H3PO4 ·0.5H2O provided 1 of the same
quality as that provided by the former process.
Experimental Section
General. All materials were purchased from commercial
suppliers and used without further purification. Melting points
were recorded on a Bu¨chi B-540 micromelting apparatus and
were uncorrected. IR spectra were recorded on a Horiba FT-
210 spectrophotometer. NMR spectra were run at 300 MHz
(1H) and 75 MHz (13C) on a Bruker DPX-300 spectrometer.
Chemical shifts are reported as δ values using tetramethylsilane
as an internal standard, and coupling constants (J) are given in
Hz. The following abbreviations are used: s ) singlet, d )
doublet, t ) triplet, m ) multiplet, and br ) broad. Optical
rotation values were recorded on a JASCO DIP-370 polarimeter
under standard conditions. HPLC analyses were performed with
Hitachi L-7000. Detection was effected with an ultraviolet
absorption photometer (wavelength 254 nm). Purity was
determined by HPLC and presented as an area percentage of
the compound peak relative to the total area of all the peaks
integrated. All compounds were judged to be of greater than
95% purity on the basis of 1H NMR and HPLC analyses. The
microanalyses and mass spectral analyses were carried out at
Takeda Analytical Research Laboratories, Ltd.
Synthesis of 2-(5-Chloro-2-nitrobenzyl)propane-1,3-diol
(7). To a solution of 5 (39.8 kg, 214.5 mol) and diethyl malonate
(34.4 kg, 214.8 mol) in acetic anhydride (86.8 kg, 850.2 mol)
was added sodium acetate (26.4 kg, 321.8 mol). The mixture
was stirred at 90-95 °C for 5.5 h and then allowed to cool to
room temperature. After adding water (119.4 kg), the mixture
was stirred at room temperature for 1 h. Then water (119.4 kg)
and toluene (206.6 kg) were added, and the layers were
separated. The aqueous layer was extracted with toluene (206.6
(23) The enantiomeric excess of 1 was determined by HPLC. HPLC
conditions: column, Chiralcel OD-RH (4.6 mm × 150 mm); mobile
phase, 0.05 M aq CH3CO2NH4/MeCN (40:60); flow rate, 1.0 mL/
min; column temperature, 40 °C; detection, UV 250 nm.
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1115

1
cm-1; MS (ESI) m/z 424 (M + Na)+; H NMR (300 MHz,
H, 7.11; N, 8.24; S, 6.29; Cl, 6.95. Found: C, 58.72; H, 7.03;
N, 8.13; S, 6.34, Cl, 6.72.
CDCl3) δ 2.56-2.64 (m, 1H), 3.04-3.08 (m, 8H), 4.23-4.36
(m, 4H), 7.42 (m, 1H), 8.03 (d, J ) 8.1 Hz, 1H); 13C NMR (75
MHz, CDCl3) δ 31.1, 37.5, 39.1, 67.4, 127.2, 128.7, 133.0,
135.1, 140.1, 147.4; Anal. Calcd for C12H16NO8S2Cl: C, 35.87;
H, 4.01; N, 3.49; S, 15.96; Cl, 8.82. Found: C, 35.85; H, 3.79;
N, 3.34; S, 16.01; Cl, 8.55.
Synthesis of 1-[(3S)-6-Chloro-1,2,3,4-tetrahydroquinolin-
3-yl]-N,N-dimethylmethanamine ((S)-2). A suspension of (S)-
2·(R)-9 (8.40 kg, 16.5 mol, 89.4% de) in 0.5 N sodium
hydroxide aqueous solution (128.5 kg) was stirred at room
temperature for 0.5 h, and then extracted with toluene (109.0
kg). The aqueous layer was extracted with toluene (36.3 kg).
The combined organic layers were washed with water (2 ×
42.0 kg) and concentrated in Vacuo. Then 1:2 toluene/n-hexane
(9.2 kg) was added to the residue. The resultant mixture was
heated and stirred at 75 °C until clear solution was obtained.
Then the solution was allowed to cool to room temperature and
stirred for 20 h. The resultant precipitate was collected by
filtration, washed with 1:2 toluene/n-hexane (2.3 kg), and dried
in Vacuo to give crude-(S)-2 (3.14 kg, 14.0 mol, 98.6% ee, 85%
yield) as a white crystalline powder. A suspension of crude-
(S)-2 (3.10 kg, 13.8 mol, 98.6% ee) in 1:2 toluene/n-hexane
(6.7 kg) was heated and stirred at 80 °C until clear solution
was obtained. Then the mixture was allowed to cool to room
temperature and stirred for 3 h. The resultant precipitate was
collected by filtration, washed with 1:2 toluene/n-hexane (2.4
kg), and dried in Vacuo to give (S)-2 (2.81 kg, 12.5 mol,
> 99.9% ee, 77% yield (based on (S)-2 ·(R)-9 used)) as a white
crystalline powder. Mp 115-116 °C; IR (KBr) ν 3271, 1604,
Synthesis of 1-(6-Chloro-1,2,3,4-tetrahydroquinolin-3-yl)-
N,N-dimethylmethanamine (rac-2). To a solution of 8 (58.7
kg, 146.1 mol) and acetic acid (307.7 kg) in THF (260.8 kg)
was added zinc dust (95.6 kg, 1462.0 mol) at 0-5 °C. The
mixture was stirred at 0 °C for 1 h and at 60 °C for 3 h. After
cooling to room temperature, the precipitate was filtered off
and washed with THF (260.8 kg). The combined filtrates were
concentrated in Vacuo. Ethyl acetate (264.6 kg) was added to
the residue. The resultant solution was washed with water (3
× 293.4 kg), 5% aq NaHCO3 (308.0 kg), and 5% aq NaCl
(308.0 kg), and then concentrated in Vacuo. Then DMSO (64.6
kg) was added to the residue, and the resultant solution was
concentrated in Vacuo. Then DMSO (64.6 kg) and 50% aq
dimethylamine (252.3 kg, 2798.4 mol) were successively added
dropwise, and the resultant mixture was stirred at 60 °C for
5 h. After the addition of water (293.4 kg), the mixture was
stirred at room temperature for 1 h and at 0 °C for 5 h. The
resultant precipitate was collected by filtration, washed with
water (293.4 kg), and dried in Vacuo to give rac-2 (25.1 kg,
111.7 mol, 77% yield) as a white crystalline powder. Mp 94-95
°C; IR (KBr) ν 3246, 1604, 1495, 1304, 806 cm-1; MS (ESI)
m/z 225 (M + H)+; 1H NMR (300 MHz, CDCl3) δ 2.11-2.26
(m, 9H), 2.43 (dd, J ) 8.7, 16.3 Hz, 1H), 2.81 (dd, J ) 4.4,
16.1 Hz, 1H), 2.97 (t, J ) 9.7 Hz, 1H), 3.38 (dd, J ) 1.4, 11.2
Hz, 1H), 3.86 (br s, 1H), 6.39 (d, J ) 7.9 Hz, 1H), 6.91 (d, J
) 7.8 Hz, 1H), 6.92 (s, 1H); 13C NMR (75 MHz, CDCl3) δ
30.2, 31.7, 45.4, 46.1, 63.1, 114.9, 121.3, 122.0, 126.6, 129.3,
143.2; Anal. Calcd for C12H17N2Cl: C, 64.13; H, 7.62; N, 12.47;
Cl 15.78. Found: C, 64.12; H, 7.60; N, 12.43; Cl, 15.67.
Synthesis of the Salt of 1-[(3S)-6-Chloro-1,2,3,4-tetrahy-
droquinolin-3-yl]-N,N-dimethylmethanamine with N-[(4-
Methylphenyl)sulfonyl]-D-leucine ((S)-2 ·(R)-9). A suspension
of rac-2 (12.0 kg, 53.4 mol) and (R)-9 (15.2 kg, 53.3 mol) in
ethyl acetate (216.5 kg) was heated to 68 °C and stirred until a
clear solution was obtained. Then the mixture was allowed to
cool to room temperature and was stirred for 18 h. The resultant
precipitate was collected by filtration, washed with ethyl acetate
(21.6 kg), and dried in Vacuo to give (S)-2·(R)-9 (8.40 kg, 16.5
mol, 89.4% de, 31% yield) as a white crystalline powder. Mp
142-143 °C; IR (ATR) ν 2361, 1497, 1164, 569, 544 cm-1;
MS (ESI) m/z 286 (M + H)+; [R]20D -0.6 (c 1.01, MeOH); 1H
NMR (300 MHz, DMSO-d6) δ 0.68 (d, J ) 6.5 Hz, 3H), 0.78
(d, J ) 6.6 Hz, 3H), 1.35 (t, J ) 7.1 Hz, 2H), 1.52-1.66 (m,
1H), 1.91-2.04 (m, 1H), 2.15-2.32 (m, 9H), 2.35 (s, 3H),
2.66-2.83 (m, 2H), 3.23 (d, J ) 9.4 Hz, 1H), 3.55 (t, J ) 7.2
Hz, 1H), 5.84 (br s, 1H), 6.41 (d, J ) 8.0 Hz, 1H), 6.83 (d, J
) 9.0 Hz, 1H), 6.85 (s, 1H), 7.33 (d, J ) 8.3 Hz, 2H), 7.63 (d,
J ) 8.2 Hz, 2H), 7.88 (br s, 1H); 13C NMR (75 MHz, DMSO-
d6) δ 21.5, 21.8, 23.2, 24.4, 29.3, 31.5, 42.0, 44.6, 45.3, 55.0,
62.1, 114.8, 118.7, 121.3, 126.6, 127.1, 129.0, 129.8, 138.8,
142.8, 144.5, 174.1; Anal. Calcd for C25H36N3O4SCl: C, 58.87;
1493, 1306, 814 cm-1; MS (ESI) m/z 225 (M + H)+; [R]20
+47.0 (c 0.96, MeOH); H NMR (300 MHz, CDCl3) δ
D
1
2.11-2.27 (m, 9H), 2.42 (dd, J ) 8.8, 16.2 Hz, 1H), 2.81 (dd,
J ) 3.2, 15.4 Hz, 1H), 2.98 (t, J ) 9.5 Hz, 1H), 3.38 (dd, J )
1.6, 11.2 Hz, 1H), 3.86 (br s, 1H), 6.39 (d, J ) 8.7 Hz, 1H),
6.91 (d, J ) 9.0 Hz, 1H), 6.92 (s, 1H); 13C NMR (75 MHz,
CDCl3) δ 30.2, 31.7, 45.4, 46.1, 63.1, 114.9, 121.3, 122.0, 126.6,
129.3, 143.2; Anal. Calcd for C12H17N2Cl: C, 64.13; H, 7.62;
N, 12.47; Cl, 15.78. Found: C, 64.16; H, 7.82; N, 12.43; Cl,
15.57.
Synthesis of (2R)-2-amino-1-[(3R)-6-chloro-3-[(dimeth-
ylamino)methyl]-3,4-dihydroquinolin-1(2H)-yl]-3-(1H-indol-
3-yl)-propan-1-one (syn-3a) (First-Generation Process). To
a solution of Fmoc-D-Trp-OH·THF (4.68 kg, 9.38 mol) and
DMF (0.28 kg) in THF (25.0 L) was added dropwise oxalyl
chloride (1.43 kg, 11.3 mol) at 0-10 °C. The mixture was
stirred at 0-10 °C for 1 h. Then a solution of (S)-2 (1.40 kg,
6.23 mol) and triethylamine (1.37 kg, 13.5 mol) in THF (12.5
L) was added dropwise, and the mixture was stirred at 0-10
°C for 3 h. After adding 5% aq NaHCO3 (18.8 L), the mixture
was allowed to warm to room temperature and was extracted
with ethyl acetate (18.8 L). The organic layer was washed with
water (18.8 L) and concentrated in Vacuo. Methanol (75.0 L)
and piperidine (3.77 kg, 44.3 mol) were successively added to
the residue. After stirring at room temperature for 16.5 h, the
precipitate was filtered off and washed with methanol (7.5 L).
The combined filtrates were concentrated in Vacuo, and then
methanol (25.0 L) was added to the residue. After stirring at
room temperature for 1 h, the precipitate was filtered off and
washed with methanol (6.3 L). The combined filtrates were
charged onto the column that was filled with DIAION WK100,
which had been conditioned according to the method described
in a footnote.22 The crude mixture was successively eluted by
1116
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

methanol (1125 L), 0.5% aq NaCl/methanol (1125 L, 20:80),
and 5 M aq NaOH/methanol (1125 L, 20:80). The combined
fractions containing syn-3a were adjusted to pH 7 by adding 2
M hydrochloric acid and then were concentrated in Vacuo. Ethyl
acetate (188 L) and sodium carbonate (18.8 kg) were added to
the residue, and the layers were separated. The organic layer
was washed with water (2 × 63.0 L) and then concentrated in
Vacuo to give amorphous solids containing syn-3a (1.88 kg,
4.57 mol, 73% yield (based on (S)-2 used)).
0.01(5), indicates the correct absolute configuration. Further
details of the X-ray structure data are available on request from
the Cambridge Crystallographic Data Centre (deposition number
CCDC 761513).
N-((2R)-1-{(3R)-6-chloro-3-[(dimethylamino)methyl]-3,4-
dihydroquinolin-1(2H)-yl}-3-(1H-indol-3-yl)-1-oxopropan-2-
yl)-1-[(1-methyl-1H-indol-2-yl)carbonyl]piperidine-4-car-
boxamide (1). A mixture of syn-3a ·0.25H3PO4 ·0.5H2O (1.40
kg, 3.15 mol), 4 (1.08 kg, 3.77 mol), EDC·HCl (0.73 kg), and
HOBt (0.58 kg) in DMF (14.0 L) was stirred at room
temperature for 2 h. Then the mixture was added to ethyl acetate
(18.0 L) at 10 °C. After adding DMF (2.0 L) and ethyl acetate
(10.0 L), the mixture was washed with 5% aq K2CO3 (2 ×
14.0 L) and 10% aq NaCl (14.0 L) below 20 °C. Then activated
carbon (0.14 kg) was added to the organic layer. After stirring
at room temperature for 10 min, the activated carbon was filtered
off and washed with ethyl acetate (4.2 L). The combined filtrates
were concentrated in Vacuo. Ethanol (4.2 L) was added to the
residue, and the resultant solution was concentrated in Vacuo.
Then ethanol (12.0 L) was added to the residue, and the resultant
solution was added dropwise to water (70.0 L) and rinsed with
ethanol (2.0 L). After stirring at room temperature for 1 h, the
resultant precipitate was collected by filtration, washed with
water (7.0 L), and dried in Vacuo to give 1 (1.89 kg, 2.75 mol,
87% yield, MW. 688.26, as 0.5 hydrate) as a white amorphous
powder. Mp 135-145 °C; IR (ATR) ν 1630, 1442, 1270, 1214,
739 cm-1; MS (ESI) m/z 679 (M + H)+; [R]20D -151.4 (c 1.00,
MeOH); 1H NMR (500 MHz, DMSO-d6) δ 1.42-1.59 (m, 2H),
1.61-1.91 (m, 3H), 1.94-2.11 (m, 8H), 2.38-2.68 (m, 5H),
2.75-3.20 (m, 4H), 3.40-3.55 (m, 1H), 3.74 (s, 3H), 3.80-4.70
(m, 2H), 5.16 (br s, 1H), 6.62 (s, 1H), 6.79 (br s, 1H), 6.90-7.35
(m, 7H), 7.43-7.66 (m, 3H), 8.36 (br s, 1H), 10.77 (br s, 1H);
13C NMR (125 MHz, DMSO-d6) δ 28.1, 30.2, 30.6, 41.2, 45.3,
47.1, 50.8, 63.5, 101.8, 109.1, 110.2, 111.2, 117.6, 118.0, 119.8,
120.7, 121.0, 122.6, 123.7, 125.7, 126.0, 126.7, 127.3, 132.4,
135.9, 137.1, 137.4, 161.8, 172.0, 173.8; Anal. Calcd for
C39H43N6O3Cl ·0.5H2O: C, 68.06; H, 6.44; N, 12.21. Found:
C, 68.36; H, 6.28; N, 12.21.
Synthesis of syn-3a·0.25H3PO4 ·0.5H2O (Second-Genera-
tion Process). A methanol solution containing syn-3a (2.18 kg,
5.30 mol) was prepared from (S)-2 (1.40 kg, 6.23 mol) by the
same procedure as the first-generation process without the CEC
and the successive separation in 85% yield. A part of the
methanol solution containing syn-3a (37.1 g, 90.3 mmol) was
concentrated in Vacuo. Ethyl acetate (200 mL) and water (100
mL) were added to the residue. The aqueous layer was adjusted
to pH 6 by adding 2 M hydrochloric acid (200 mL), and then
the layers were separated. The organic layer was extracted with
water (50 mL). After adding ethyl acetate (200 mL) and Na2CO3
(50 g) to the combined aqueous layer, the layers were separated.
The organic layer was washed with 5% aq NaCl (100 mL) and
concentrated in Vacuo to give crude-syn-3a (42.7 g). Phosphoric
acid (0.60 mL, 0.85 M) was added to a solution of the crude-
syn-3a (1.00 g) in ethanol (2.0 mL) and water (2.0 mL). After
stirring at room temperature for 0.5 h, water (4.0 mL) was added
dropwise. The mixture was allowed to cool to 0 °C and stirred
for 1 h. The resultant precipitate was collected by filtration,
washed with 10% aq ethanol (1.0 mL), and dried in Vacuo to
give crude-syn-3a ·0.25H3PO4 ·0.5H2O (858 mg). After the
crude-syn-3a ·0.25H3PO4 ·0.5H2O (500 mg) was suspended in
ethyl acetate/n-hexane (1.5 mL/0.5 mL) at room temperature
for 0.5 h, the precipitate was collected by filtration, washed with
n-hexane (1.0 mL), and dried in Vacuo to give syn-
3a ·0.25H3PO4 ·0.5H2O (430 mg, 0.978 mmol, 68% yield (based
on (S)-2 used)) as a white crystalline powder. Mp 159-166
°C; IR(KBr) ν 1658, 1485, 1090, 1043, 744 cm-1; MS (FAB)
m/z 411 (M + H)+; [R]20D +202.1 (c 1.00, MeOH.); 1H NMR
(300 MHz, DMSO-d6) δ 1.69-1.94 (m, 3H), 2.05 (s, 7H),
2.20-3.50 (m, 5H), 4.08-4.22 (m, 1H), 6.76-6.90 (m, 1H),
6.91-7.18 (m, 5H), 7.29 (d J ) 8,1 Hz, 1H), 10.8 (s, 1H); 13C
NMR (75 MHz, DMSO-d6) δ 30.2, 32.5, 33.0, 45.5, 46.9, 52.5,
63.2, 110.1, 111.4, 118.2, 118.3, 120.9, 123.6, 125.7, 127.2,
127.6, 128.8, 136.2, 137.7, 174.9; Anal. Calcd for
C23H27N4OCl ·0.25H3PO4 ·0.5H2O: C, 62.16; H, 6.52; N, 12.61;
Cl, 7.98; P, 1.74. Found: C, 62.16; H, 6.40; N, 12.47; Cl, 7.93;
P, 1.77. Crystal data for syn-3a ·0.25H3PO4 ·0.5H2O:
C23H27ClN4O, 0.25H3PO4, 0.5H2O, tetragonal, I4 (#79), a )
22.700 (3) Å, c ) 8.762(3) Å, V ) 4514.8(17) Å3, Z ) 8, Dcalc
) 1.308 g/cm3, R ) 0.056, Rw ) 0.180. The Flack parameter,24
Acknowledgment
We thank Dr. Masahiro Kajino, Dr. Shokyo Miki, Dr.
Tomomi Ikemoto, Dr. Masahiro Mizuno, Dr, Jun-ichi Kawaka-
mi, and Dr. Keisuke Majima for helpful discussions, Ms. Keiko
Higashikawa for help with the single-crystal X-ray analysis,
Ms. Kumiko Doen for help with chiral HPLC analysis, and
Mr. Ryou Sugiyama for help with DSC analysis.
Received for review March 28, 2010.
OP100084R
(24) Flack parameter: Flack, H. D. Acta Crystallogr., Sect. A 1983, 39,
876.
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1117