Synthesis of (S)-3-(N-methylamino)-1-(2-thienyl)propan-1-ol: Revisiting Eli Lilly's resolution-racemization-recycle synthesis of duloxetine for its robust processes

Article abstract of DOI:10.1021/op060118l

(±)-3-(N,N-Dimethylamino)-1-(2-thienyl)propan-1-ol (6), prepared from 2-acetylthiophene (4) in a two-step overall yield of 79%, is resolved into (S)-6 of 93% ee as its diastereomeric salt (8) with (S)-mandetic acid (7) according to Eli Lilly's procedures developed for the resolution-racemization- recycle (RRR) synthesis of duloxetine (2) with some modifications in terms of practicality. On its liberation from 8, (S)-6 undergoes N-demethylative ethyl carbamate formation in two discrete but successive steps in an overall yield of 87% from 8: (1) O-ethyl carbonate formation and (2) ethyl carbamate formation with concomitant loss of the N-methyl group. Alkaline hydrolysis then affords (S)-3-(N-methylamino)-1-(2-thienyl)propan-1-ol (1) of 100% ee, an alleged penultimate precursor to duloxetine (2), in 75% yield after a single recrystallization from ethylcyclohexane. In the overall process thus developed, PhMe is substituted successfully for t-BuOMe, a solvent that has been used favorably in Eli Lilly's original RRR synthesis of 2.

Full text of DOI:10.1021/op060118l

Organic Process Research & Development 2006, 10, 905−913
Synthesis of (S)-3-(N-Methylamino)-1-(2-thienyl)propan-1-ol: Revisiting Eli
Lilly’s Resolution-Racemization-Recycle Synthesis of Duloxetine for Its
Robust Processes
Yoshito Fujima, Masaya Ikunaka,* Toru Inoue, and Jun Matsumoto
Research & DeVelopment Center, Nagase & Co., Ltd. 2-2-3 Murotani, Nishi-ku, Kobe, 651-2241 Japan
Abstract:
(()-3-(N,N-Dimethylamino)-1-(2-thienyl)propan-1-ol (6), pre-
pared from 2-acetylthiophene (4) in a two-step overall yield of
79%, is resolved into (S)-6 of 93% ee as its diastereomeric salt
(8) with (S)-mandelic acid (7) according to Eli Lilly’s procedures
developed for the resolution-racemization-recycle (RRR)
synthesis of duloxetine (2) with some modifications in terms of
practicality. On its liberation from 8, (S)-6 undergoes N-
demethylative ethyl carbamate formation in two discrete but
successive steps in an overall yield of 87% from 8: (1) O-ethyl
carbonate formation and (2) ethyl carbamate formation with
concomitant loss of the N-methyl group. Alkaline hydrolysis
then affords (S)-3-(N-methylamino)-1-(2-thienyl)propan-1-ol (1)
of 100% ee, an alleged penultimate precursor to duloxetine (2),
in 75% yield after a single recrystallization from ethylcyclo-
hexane. In the overall process thus developed, PhMe is
substituted successfully for t-BuOMe, a solvent that has been
used favorably in Eli Lilly’s original RRR synthesis of 2.
Figure 1. Structures of (S)-3-(N-methylamino)-1-(2-thienyl)-
propan-1-ol (1), duloxetine (2), and fluoxetine (3).
technologies has been explored to date for scalable processes
that should provide 2 of high stereochemical purity;5 and
among other things, resolution via diastereomeric salt forma-
tion seemed the most industrially viable as implied in
Weigel’s review.1
The resolution-racemization-recycle (RRR) synthesis of
2 developed at Eli Lilly starts with the Mannich reaction
(Me2NH‚HCl, CH2O, HCl, i-PrOH) of 2-acetylthiophene (4)
(Scheme 1).1,5b The â-aminoketone freed from Mannich
product (5) by base treatment (NaOH, EtOH; pH 11-12) is
subjected to NaBH4-mediated reduction. After acidic workup
(HCl; pH 1-1.5) cleaving both B-O and B-N linkages,
basification (pH 12) affords (()-(N,N-dimethyl)aminoalcohol
(6) as a free base,5b which is extracted into t-BuOMe
(MTBE). To the MTBE solution is added (S)-mandelic acid
(7) (0.45 equiv) dissolved in EtOH. The resulting slurry
[MTBE/EtOH (9.8:1)] is heated to reflux and then cooled
to ambient temperature to allow (S)-7 to form a diastereo-
meric salt with (S)-6, during which (R)-6 is left unaffected
and remains dissolved in the solution. The precipitated salt
(8) [(S)-6‚(S)-7] is collected by filtration and treated with
aqueous NaOH solution to liberate (S)-6. In the meantime,
the MTBE solution of (R)-6 recovered as the filtrate (mother
liquor) is treated with HCl to racemize (R)-6 for another
round of the resolution (recycling) as outlined in Scheme
1.1
Introduction
This communication will deal with the synthetic processes
for (S)-3-(N-methylamino)-1-(2-thienyl)propan-1-ol (1) which
were developed with modifications to Lilly’s resolution-
racemization-recycle (RRR) synthesis of duloxetine (2)
(Figure 1).1
Being a potent dual inhibitor of serotonin and noradrenalin
reuptake, 2 has therapeutic potency to treat stress urinary
incontinence as well as major depressive disorder.2 In
contrast, its ancestor, fluoxetine (3)3 which represents a
selective serotonin reuptake inhibitor (SSRI), is prescribed
only as an antidepressant (Figure 1). In addition, contrary
to 3 that was approved as a racemate,4 2 was supposed to
reach the therapeutic market as a single enantiomer from
the outset of its development. Thus, a wide range of chiral
* To whom correspondence should be addressed. Masaya Ikunaka. Tele-
phone: +81 78 992 3162. Fax: +81 78 992 1050. E-mail: masaya.ikunaka@
nagase.co.jp.
(1) Astleford, B. A.; Weigel, L. O. Resolution Versus Stereoselective Synthesis
in Drug Development: Some Case Studies. In Chirality in Industry II:
DeVelopments in the Commercial Manufacture and Applications of Optically
ActiVe Compounds; Collins, A. N., Sheldrake, G. N., Crosby, J., Eds.; John
Wiley & Sons: Chichester, 1997; pp 99-117.
(2) Sorbera, L. A.; Castan˜er, R. M.; Castan˜er, J. Drugs Future 2000, 25, 907.
(3) (a) Li, J. J.; Johnson, D. S.; Sliskovic, D. R.; Roth. B. D. Antidepressants.
In Contemporary Drug Synthesis; Wiley-Interscience: Hoboken, 2004;
pp125-147. (b) Wirth, D. D.; Miller, M. S.; Boini, S. K.; Koenig, T. M.
Org. Process Res. DeV. 2000, 4, 513.
(4) (a) Hilborn, J. W.; Lu, Z.-H.; Jurgens, A. R.; Fang, Q. K.; Byers, P.; Wald,
S. A.; Senanayake, C. H. Tetrahedron Lett. 2001, 42, 8919. (b) Liu, H.-L.;
Hoff, B. H.; Anthonsen, T. J. Chem. Soc., Perkin Trans. I 2000, 1767.
(5) (a) Deeter, J.; Frazier, J.; Staten, G.; Staszak, M.; Weigel, L. Tetrahedron
Lett. 1990, 31, 7101. (b) Berglund, R. A. (Eli Lilly & Co.). U.S. Patent
5,362,886, 1994. (c) Liu, H.; Hoff, B. H.; Anthonsen, T. Chirality 2000,
12, 26. (c) Kamal, A.; Khanna, G. B. R.; Ramu, R.; Krishnaji, T.
Tetrahedron Lett. 2003, 44, 4783. (d) Wang, G.; Liu, X.; Zhao, G.
Tetrahedron: Asymmetry 2005, 16, 1873. (e) Iwakura, K.; Azumai, T.;
Sakahigashi, S. (Sumitomo Chemical Co., Ltd. and Sumitomo Seika
Chemicals Co., Ltd.). Jpn. Kokai Tokkyo Koho 2004-344067, 2004.
10.1021/op060118l CCC: $33.50 © 2006 American Chemical Society
Published on Web 09/15/2006
Vol. 10, No. 5, 2006 / Organic Process Research & Development
•
905

Scheme 1. Eli Lilly’s RRR synthesis of
(S)-3-(N,N-dimethylamino)-1-(2-thienyl)propan-1-ol (6)
provides duloxetine hydrochloride (2‚HCl) as a white solid
in an overall yield of 32% from (S)-6.
Recently, it was suggested that (S)-3-(N-methylamino)-
1-(2-thienyl)propan-1-ol (1) (Figure 1) should represent
another penultimate precursor to 2 in place of 10.6 Hence,
as part of our process development program featuring
resolution via diastereomeric salt formation,7 we chose to
access (S)-6 following the RRR tactics that had originated
in Eli Lilly’s protocol and then subject it to N-demethylation
via carbamate formation in having a scalable access to (S)-1
as will be discussed below in full detail.
Results and Discussion
Preparation of (()-Alcohol (6). The Mannich reaction
with 4 was first conducted under the same conditions as
reported from Eli Lilly (Scheme 1). When a mixture of 4
(1.0 equiv), Me2NH‚HCl (1.26 equiv), paraformaldehyde
(1.49 equiv), HCl (0.13 equiv), and i-PrOH (2.4 v/w) was
heated to reflux all at once, there took place abrupt
precipitation of the crystalline Mannich product (5) with
vigorous bumping, which raised safety concerns. Thus, to
prevent this unfavorable phenomenon, two countermeasures
were taken: (1) To a solution of Me2NH‚HCl (1.25 equiv),
paraformaldehyde (1.20 equiv) and 35% aqueous HCl
solution (0.25 equiv) in i-PrOH (2.0 v/w) was added 4 in
two halves; (2) prior to the addition of the second half of 4,
the reaction mixture was seeded with 5 to induce partial
crystallization deliberately. As a result, these two modifica-
tions allowed 5 to be obtained as white crystals in 88% yield
uneventfully in an industrially viable manner (Scheme 3).
The free â-N,N-dimethylaminoketone liberated from 5
was reduced with NaBH4 (0.50 equiv) in EtOH according
to Eli Lilly’s procedures as outlined in Scheme 1;5b however,
the acidic workup (HCl; pH 1-1.5) aimed at hydrolyzing
both N-B and O-B bonds was inevitably accompanied by
ethyl ether (13) being generated in a significant amount,
which should arise from nucleophilic addition of EtOH to a
cationic species that could be represented as two limiting
structures, 11 and 12 (Scheme 3). To avoid such ether
formation, the NaBH4-mediated reduction was conducted in
an aqueous medium free of EtOH. The Mannich product (5)
was added directly to an alkaline suspension of NaBH4 (0.4
equiv) in an aqueous solution of NaOH (1.1 equiv). The
reaction was then quenched with 35% aqueous HCl solution
(1.65 equiv) to destroy all the boron complexes. The mixture
was basified with 48% aqueous NaOH solution to liberate
free (()-(N,N-dimethyl)aminoalcohol (6), which was then
extracted into PhMe. Eventually, crystallization from eth-
ylcyclohexane (ECH) provided (()-6 as a white solid in 90%
yield.
Scheme 2. Eli Lilly’s assemblage of 2‚HCl from (S)-6
When (S)-6 is treated with NaH (1.0 equiv) in the presence
of PhCO2K (0.1 equiv) in DMSO,5b it is allowed to
participate in aromatic nucleophilic substitution on 1-fluo-
ronaphthalene (9) (1.2 equiv), as depicted in Scheme 2. The
coupled product is then isolated as its phosphate salt (10) of
91% ee in 79.6% yield from (S)-6. On basification under
the biphasic conditions (NH3, H2O, PhMe), the free amine
liberated from 10 is taken up into PhMe. To the PhMe
solution is added i-Pr2EtN (0.10 equiv) followed by PhO-
COCl (1.25 equiv), and the resulting homogeneous mixture
is heated to 55 °C to implement N-demethylation. The
resulting phenyl carbamate (11) is finally subjected to
alkaline hydrolysis (NaOH, H2O, DMSO) without isolation.
Aqueous workup (acidification to pH 5.0-5.5 with AcOH,
washing with n-hexane, basification to pH 10.5 with 50%
aqueous NaOH solution, extraction with AcOEt) followed
by salt formation with HCl in the AcOEt solution eventually
Resolution of (()-6 into (S)-6 as its Diastereomeric Salt
with (S)-Mandelic Acid (7). The resolution procedures
established at Eli Lilly1,5b were reexamined for further
improvement. (()-Alcohol (6) was combined with (S)-
mandelic acid (7) (0.45 equiv) in a range of solvents listed
(6) Sakai, K.; Sakurai, R.; Yuzawa, A.; Kobayashi, Y.; Saigo, K. Tetrahe-
dron: Asymmetry 2003, 14, 1631.
(7) Hirayama, Y.; Ikunaka, M.; Matsumoto, J. Org. Process Res. DeV. 2005,
9, 30.
906
•
Vol. 10, No. 5, 2006 / Organic Process Research & Development

Scheme 3. Preparation of (()-6 free from ether byproduct
(13)
(8) (entry 6, Table 1). Eventually, under these conditions,
the precipitated salt was isolated in 41% yield, (S)-6
contained in it being of 93.0% ee; and, what was better, the
enantiomeric purity of (S)-6 remained unchanged even after
the salt had been suspended in the mother liquor containing
(R)-6 at ambient temperature for a day.
As a result, the solvent system of MTBE-EtOH (9.8:1)
that had been employed for the diastereomeric salt formation
at Eli Lilly was replaced successfully with that of PhMe-
MeOH (40:1), which was more economical and environ-
mentally benign: when (()-6 was combined with (S)-7 (0.45
equiv) in PhMe (10 v/w)-MeOH (0.25 v/w), (S)-6 was
allowed to form a diastereomeric salt with (S)-7 in a highly
reproducible manner to give (8) in 41% yield with (S)-6 of
93% ee being contained in it, as summarized in Scheme 4.
While the separated salt (8) was carried forward to the
N-demethylation stage, the filtrate (mother liquor) was
concentrated to recover (R)-6, which should be racemized
for another round of the resolution, as discussed below.
Acid-Catalyzed Racemization of (R)-Alcohol (6). Ra-
cemization of the off-enantiomer, (R)-6, was also attempted
according to the procedures devised at Eli Lilly.1 When (R)-6
of 70.8% ee at a concentration of 0.67 M was treated with
3 M aqueous HCl solution (3.0 equiv) at room temperature
for 5 h, its enantiomeric purity declined to 1.8% ee. However,
two kinds of byproducts were found to be generated in the
course of the acid-induced racemization, as shown by HPLC
analysis [YMC-Pack ODS-AM, i.d. 6.0 mm × 150 mm;
H2O/MeCN/CF3CO2H (80:20:0.01); UV 254 nm; for more
detail, see the analytical condition (3) in the Experimental
Section]: tR 2.8 min (10.2 area %) and tR 4.8 min (6.7 area
%).
The slower-eluting byproduct (tR 4.8 min) could be
isolated using preparative TLC, and its structure was
identified as olefin (14) by 1H NMR and GC-MS analysis,
which should arise from acid-catalyzed dehydration (Scheme
4). As regards the faster-eluting one (tR 2.8 min), which was
difficult to isolate in a pure state by silica gel chromatog-
raphy, it was deduced from LC-MS (ESI) that its structure
should be bis(2-thienyl)methine (15) based on [M + H]+ )
353 (C18H28N2OS2); the plausible mechanism of the forma-
tion of 15 should involve electrophilic aromatic substitution
as outlined in Scheme 4.
in Table 1 to precipitate diastereomeric salt (8), which was
analyzed for the enantiomeric purity of (S)-6 contained in it
by HPLC [column, Daicel OD i.d. 4.6 mm × 250 mm;
eluent: n-hexane/i-PrOH/Et2NH (96.5:3.5:0.1); for more
detail, see the analytical condition (2) in the Experimental
Section]. When 8 was formed in MTBE doped with a protic
solvent such as EtOH (10%; entry 1, Table 1) and H2O (1%;
entry 2, Table 1), efficient chiral separation was reproduced
as reported from Eli Lilly.
MTBE being a solvent that has posed environmental
concerns, other solvents of similar polarity, such as Bu2O
and i-PrOAc, were applied to the diastereomeric salt forma-
tion. While the use of Bu2O/H2O (100:1) ended in less
efficient resolution (entry 3, Table 1), that of i-PrOAc (entry
4, Table 1) enabled chiral separation as efficient as that
recorded in the literature from Eli Lilly.5b
To prevent both 14 and 15 from being formed or at least
to put their formation under stringent control, concentrations
of (R)-6 and amounts of HCl applied were varied for the
best combination since both factors seemed to affect the
chemical fate of 11 and/or 12. When (R)-6 (70.8% ee; 0.67
M) was treated with 2 M aqueous HCl solution (3.0 equiv)
at 21°C, it took 8 h until its enantiomeric purity dropped to
below 5% ee, with 14 and 15 being produced in 7.1 and
12.7 area %, respectively, as shown by HPLC (entry 1, Table
2). When the concentration of (R)-6 (70.8% ee) was kept
constant at 0.67 M, increase in the amount of HCl (5 M)
from 3.0 to 8.0 equiv accelerated the reaction rate 4 times,
but it also raised the amounts of 14 and 15 to 18.9 and 37.8
area %, respectively (entry 2, Table 2).
The diastereomeric salt formation in question was next
attempted in a solvent of low polarity. When 8 was allowed
to precipitate from PhMe, its yield amounted to 43%, the
enantiomeric purity of (S)-6 contained in it being 94.4% ee
(entry 5, Table 1). However, the enantiomeric purity of (S)-6
contained in 8 suffered a decrease from 94.4% ee to 86.1%
ee when the precipitated salt (8) was left suspended in the
mother liquor [the PhMe solution of (R)-6] at 25 °C for 12
h.
It was then indicated from experimentation that diaster-
eomeric salt (8) would gain the stereochemical stability in
PhMe when it was doped with small amounts of a protic
solvent; hence, PhMe doped with MeOH (2.5%) turned out
to be the medium of choice in forming diastereomeric salt
Vol. 10, No. 5, 2006 / Organic Process Research & Development
•
907

Table 1. Solvent screen for the diastereomeric salt formation of (S)-6 with (S)-mandelic acid (7)
entry
solvent (v/w)a
yield (%) of 8
ee (%) of (S)-6 in 8b
resolution efficiencyc
1
2
3
4
5
6
MTBE (10), EtOH (1)
MTBE (10), H2O (0.1)
Bu2O (10), H2O (0.1)
AcOPri (10)
PhMe (10)
PhMe (10), MeOH (0.25)
41.8
42.4
33.4
41.7
43.4
41.0
87.2
91.0
76.4
91.2
94.4
93.0
36.4
38.6
25.5
38.0
41.0
38.1
a A ratio of the solvent volume (mL) relative to the weight (g) of (()-6. b Determined by HPLC under the analytical condition (2) detailed in the Experimental
Section. c Calculated according to the following equation: [yield (%) of 8] × [ee (%) of (S)-6 in 8] × 0.01
Scheme 4. Resolution of (()-6 via diastereomeric salt
formation with (S)-mandelic acid (7) and acid-catalyzed
racemization of (R)-6 accompanied by dehydrative formation
of byproducts (14) and (15)
Table 2. Conditions for the acid-promoted racemization of
(R)-6
HCl
[(R)-6]
(M)
(equiv)
temp time
14
15
entry
1
[HCl] (M) (°C) (h) (area %)a (area %)a
3
2
0.67
0.67
0.34
0.34
0.34
0.17
21
21
21
9
8
2
7.1
18.9
5.6
12.7
37.8
7.5
8
5
2
3
4
5
6
6
2
4
6
2
24
8
2.1
7.7
4.5
1.5
21
21
3.8
6.0
12
2
4
6.0
6.0
a Determined by HPLC under the analytical condition (3) detailed in the
Experimental Section.
enantiomeric purity of (R)-6 to decline to below 5% ee where
formation of 14 was further lessened to 2.1 area % but that
of 15 remained almost unchanged as 7.7 area % (entry 4,
Table 2).
The acid-catalyzed racemization was next attempted using
HCl in less amounts and at lower concentrations. In the
presence of 1.5 M aqueous HCl solution (4.5 equiv), the
enantiomeric purity of (R)-6 dropped from 70.8% ee to below
5% ee at 21 °C in 8 h during which accumulation of 14 and
15 could be suppressed to 3.8 and 6.0 area %, respectively
(entry 5, Table 2). By comparison, (R)-6 (70.8% ee) at a
more dilute concentration of 0.17 M was treated with as
much as 12 equiv HCl (2 M aqueous solution) at 21 °C where
the enantiomeric purity of (R)-6 dropped to below 5% ee in
4 h with a slight increase in the amount of 14 (6.0 area %),
although that of 15 (6.0 area %) was almost unaffected when
compared with entry 5 in Table 2 (entry 6, Table 2).
In the event, the racemization was best accomplished with
(R)-6 (70.8% ee) at its concentration of 0.34 M by the action
of 1.5 M aqueous HCl solution (4.5 equiv) at 21 °C, whereby
formation of 14 and 15 could be suppressed to not more
than 3.8 and 6.0 area %, respectively, both being the lowest
level ever attained (entry 5, Table 2).
It was assumed that since 15 was produced by a
bimolecular reaction, its amounts could be reduced by
decreasing concentrations of (R)-6. When (R)-6 (70.8% ee)
at a concentration of 0.34 M was treated with 2 M aqueous
HCl solution (6.0 equiv) at 21 °C, its enantiomeric purity
dropped to below 5% ee in 4 h with formation of 14 and 15
being reduced from 7.1 area % (entry 1, Table 2) to 5.6 area
% and from 12.7 area % (entry 1, Table 2) to 7.5 area %,
respectively (entry 3, Table 2).
When the reaction temperature was decreased from 21
to 9 °C, it took longer reaction time (24 h) for the
To telescope the racemization step in question into the
previous resolution step, the PhMe solution of (R)-6 that had
908
•
Vol. 10, No. 5, 2006 / Organic Process Research & Development

Scheme 5. Attempted single-step procedures for
N-demethylative ethyl carbamate formation from (S)-6
been separated from diastereomeric salt (8) by filtration was
subjected directly to the racemization conditions identified
as above. The PhMe solution of (R)-6 (63% ee) obtained as
the filtrate was washed with 1 M aqueous NaOH solution to
remove (S)-7 dissolved in it and with saturated aqueous NaCl
solution. The PhMe solution was concentrated until its
volume reached 0.42 times the original volume, whereby the
MeOH dissolved in it was removed completely; unless the
PhMe solution was freed of MeOH, the ensuing acid-
mediated racemization would suffer from methyl ether
formation as ethyl ether (13) formed in the presence of EtOH
during the acidic workup of the NaBH4-mediated reduction
of 5 (Scheme 3). To the PhMe solution thus condensed was
added 1.5 M aqueous HCl solution (4.5 equiv) and the
resulting two-phase mixture was stirred at ambient temper-
ature for 7 h. Finally, recrystallization from ECH (4 v/w)
provided (()-6 in 79-83% yield, which was free of 14 but
was contaminated with 15 in 3.0 area %.
As regards (()-6 that had been obtained from the above-
mentioned process of acid-catalyzed racemization, its reuse
was investigated. Regenerated (()-6 contaminated by 15 (3.0
area %) was combined with an equiamount of fresh (()-6
to prepare a sample of (()-6 that contained 15 in 1.5 area
%, which was then subjected to the second round of
resolution by (S)-7 [0.45 equiv relative to (()-6]. When (S)-6
thus prepared was allowed to undergo N-demethylation under
the conditions detailed below, there was produced (S)-1 of
100% ee which was contaminated by bis-nor 15, the
byproduct arising from double N-demethylation of 15, only
in 0.46 area %; the product/byproduct ratios were measured
arbitrarily in area % at 254 nm without correction based on
the absorption coefficient of each compound.
Scheme 6. Two-step procedures for N-demethylative ethyl
carbamate formation from (S)-6 and completion of the
synthesis of (S)-1
N-Demethylation of (S)-N,N-Dimethylaminoalcohol (6).
N-Demethylation of (S)-6 of 93% ee, liberated from 8 by
basification with an aqueous NaOH solution, was first
attempted using Teva’s procedures that had originally been
applied to the synthesis of 3.3b,8 A solution of (S)-6 in PhMe
was treated with ethyl chloroformate (3.0 equiv) in the
presence of NaHCO3 (1.3 equiv) with heating at reflux.8 The
HCl evolving with the progress of the reaction was neutral-
ized with solid NaHCO3, and the resulting H2O was removed
by distillation as an azeotrope with PhMe. On complete
consumption of (S)-6, aqueous workup provided (S)-N,O-
bis-ethoxycarbonylated (16) in 94% yield (Scheme 5).
However, the reaction turned out to be capricious and less
reproducible, giving rise to a complex mixture of byproducts
including dehydrated olefins, 14 and 17.
The causative agent for such byproduct formation was
assumed to be the HCl that survived the neutralization by
NaHCO3 under the heterogeneous conditions, considering
that 2-thienyl carbinol (6) was so acid-sensitive a compound
as discussed above and outlined in Schemes 3 and 4. Thus,
to prevent (S)-6 from entering into acid-promoted side
reactions, such as dehydration, its hydroxyl group was
converted to less acid-labile ethyl carbonate completely under
mild conditions prior to N-demethylative ethoxycarbonyla-
tion, which would require harsh conditions, such as elevated
reaction temperatures (Scheme 6). To a PhMe solution of
(S)-6 of 93% ee was added a PhMe solution of ethyl
chloroformate (1.3 equiv) in the presence of solid NaHCO3
(2.0 equiv) with heating at 90 °C. Aqueous workup followed
by azeotropic removal of the generated water afforded a dried
PhMe solution of (S)-18. To the PhMe solution was then
added another PhMe solution of ethyl chloroformate (2.0
(8) Schwartz, R. E.; Kaspi, G. J.; Itov, R.-L. Z.; Pilarski, H. G. (Teva
Pharmaceutical Industries Ltd.). U.S. Patent 5,225,585, 1993.
Vol. 10, No. 5, 2006 / Organic Process Research & Development
•
909

equiv) in the presence of solid NaHCO3 (0.1 equiv) with
heating at reflux to afford (S)-16 of 91.1% chemical purity
as a pale-yellow oil in 90% yield.
16 (Scheme 6). Last but not least, MTBE that had been
employed in the original RRR synthesis of 2 was replaced
with PhMe, a more favorable solvent in terms of safety and
economy, throughout the modified processes.10
Both N- and O-ethoxycarbonyl groups were cleaved from
(S)-16 by alkaline hydrolysis.3b,8 To a solution of (S)-16 in
EtOH (1.7 v/w) and H2O (2.2 v/w) was added 48% aqueous
solution of NaOH (8.5 equiv) (Scheme 6). After stirring at
80 °C for 2 h, extraction with PhMe at 50 °C furnished crude
(S)-1 as a pale-yellow solid quantitatively; its enantiomeric
purity was determined to be 93% ee by the method
mentioned below, which indeed was the same as that of
(S)-6.
Finally, recrystallization from ECH/PhMe (3:1) gave
purified (S)-1 (100 area %) as a white solid in 75% yield. It
was then attempted to determine the enantiomeric purity of
(S)-1 in a direct way, but to no avail; hence, (S)-1 thus
obtained was reverted to (S)-6, and the latter was analyzed
for enantiomeric purity. When a solution of (S)-1 in Cl(CH2)2-
Cl was treated with 37% aqueous solution of HCHO in the
presence of NaBH(OAc)3,9 reductive N-methylation pro-
ceeded via cyclic (S)-N,O-acetal (19) and eventually led to
(S)-6; the intermediacy of 19 was confirmed by its unam-
biguous preparation in which (S)-1 was treated with excess
paraformaldehyde in MTBE. Eventually, HPLC analysis
assured that (S)-N,N-dimethylamine (6) thus prepared was
of 100% ee, and as such, the enantiomeric purity of the
original (S)-1 was confirmed to be of 100% ee.
Experimental
General. Melting points were measured on an Electro-
thermal 1A8104 melting point apparatus and are uncorrected.
1H NMR spectra were recorded at 400 MHz on a Varian
UNITY-400 spectrometer in a CDCl3 solution with tetram-
ethylsilane as an internal standard. FT-IR spectra were
recorded on a Nicolet Avatar 360 FT-IR spectrometer. Mass
spectra were recorded on a Hitachi M-8000 mass spectrom-
eter (ESI). Elemental analyses were performed on an
Elementar vario EL analyzer. Optical rotations were mea-
sured on a Horiba SEPA-200 polarimeter. Thin-layer chro-
matography (TLC) was performed on Merck Kieselgel 60
plates (0.25 mm thick, art 1.05714.).
Analytical Conditions. (1) HPLC to monitor the progress
of the reduction of 5 to (()-6: column, YMC-Pack ODS-
AM, i.d. 6.0 mm × 150 mm; eluent, H2O/MeCN/trifluoro-
acetic acid (90:10:0.01); flow rate, 1.0 mL/min; column
temperature, 40 °C; detection, UV at 254 nm; tR, 5.3 min
for (()-6 and 6.0 min for 5. (2) HPLC to determine the
enantiomeric composition of 6: column, Daicel CHIRAL-
CEL OD i.d. 4.6 mm × 250 mm; eluent: n-hexane/i-PrOH/
Et2NH (96.5:3.5:0.1); flow rate, 0.7 mL/min; column tem-
perature, room temperature; detection, UV at 230 nm; tR,
15.8 min for (R)-6 and 17.3 min for (S)-6. (3) HPLC to
determine the chemical composition of 6 regenerated after
acid-catalyzed racemization of (R)-6: column, YMC-Pack
ODS-AM i.d. 6.0 mm × 150 mm; eluent, H2O/MeCN/
trifluoroacetic acid (80:20:0.01); flow rate, 1.0 mL/min;
column temperature, 40 °C; detection, UV at 254 nm; tR,
2.8 min for 15, 3.2 min for 6, and 4.8 min for 14. (4) HPLC
to monitor the progress of the reaction from (S)-6 to (S)-16
via (S)-18: column, YMC-Pack ODS-AM i.d. 6.0 mm ×
150 mm; eluent, H2O/MeCN/trifluoroacetic acid (50:50:
0.01); flow rate, 1.0 mL/min; column temperature, 40 °C;
detection, UV at 230 nm; tR, 2.2 min for 6, 2.6 min for 18
and 14.4 min for 16. (5) HPLC to monitor the progress of
the hydrolysis of (S)-16 to (S)-1 and to determine the
chemical purity of (S)-1: column, YMC-Pack ODS-AM i.d.
6.0 mm × 150 mm; eluent, H2O/MeCN/ trifluoroacetic acid
(50:50:0.01); flow rate, 1.0 mL/min; column temperature,
40 °C; detection, UV at 254 nm; tR, 2.3 min for (S)-1, 3.0
min for bis-nor 15 (arising from double N-demethylation of
15) and 14.1 min for 16. (6) GLC to monitor the progress
of the reductive N-methylation of (S)-1 to (S)-6: column,
Hicap (CBP1-M25-025), i.d. 0.25 mm × 25 m; column
temperature, 130 °C; injection temperature, 250 °C; detection
temperature, 275 °C; carrier gas, He, 90 mL/min; split ratio,
1:100; sample size, 2.0 µL; detection, FID; tR, 15.7 min for
(S)-6, 17.7 min for (S)-1, and 18.1 min for 19.
Conclusions
2-Acetylthiophene (4) was converted to (S)-3-(N-meth-
ylamino)-1-(2-thienyl)propan-1-ol (1) of 100% ee, an alleged
penultimate precursor to 2, according to Eli Lilly’s proce-
dures (Schemes 1 and 2) with modifications aimed at dealing
with the acid lability of the secondary alcohol function in
3-(N,N-dimethylamino)-1-(2-thienyl)propan-1-ol (6) in the
following three situations: (1) NaBH4-mediated reduction
of the Mannich product (5) was effected in an aqueous
medium exclusive of any alcoholic solvent, such as EtOH,
to prevent ether formation in the course of acidic workup
(Scheme 3). (2) As regards the acid-catalyzed racemization
of the off-enantiomer, (R)-6, the conditions were tuned in
terms of the substrate concentrations and the amounts of HCl
applied such that dehydrated olefin (14) and bis(2-thienyl)-
methine derivative (15) should be formed in minimum
amounts; when (R)-6 (70.8% ee; 0.34 M) was treated with
1.5 M aqueous solution of HCl (4.5 equiv) at 21 °C, the
racemization went almost to completion, giving (R)-6 of not
more than 5% ee, with 14 and 15 being formed in 3.8 and
6.0 area %, respectively (Scheme 4). (3) N-Demethylative
ethoxycarbonylation of (S)-6 to (S)-16 was conducted through
two-step discrete but successive operations to prevent the
secondary alcohol of (S)-6 from being eliminated by the
action of the HCl that adventitiously survived neutralization
by NaHCO3; (S)-6 was first converted to (S)-O-ethoxycar-
bonylated (18) at 90 °C, which was then subjected to
N-demethylative ethoxycarbonylation at 120 °C to give (S)-
(()-2-[3-(N,N-Dimethyl)aminopropionyl]thiophen Hy-
drochloride (5). To a stirred solution of 35% aqueous
(9) Abdel-Magid, A. F.; Carson, K. G; Harris, B. D.; Maryanoff, C. A.; Shah,
(10) The completed process was transferred to a contract manufacturer and
R. D. J. Org. Chem. 1996, 61, 3849.
enabled it to produce tens of kilograms of (S)-1 in a reliable manner.
910
•
Vol. 10, No. 5, 2006 / Organic Process Research & Development

solution of HCl (1.15 g, 55 mmol) in i-PrOH (55 mL) were
added paraformaldehyde (95%; 8.35 g, 264 mmol), Me2NH‚
HCl (22.43 g, 275 mmol), and 2-acetylthiophene (4) (13.88
g, 110 mmol) in sequence at room temperature. The mixture
was heated to 67-70 °C, and the stirring was continued with
heating at the same temperature range for 1 h. The
homogeneous mixture was seeded with a few crystals of 5
to induce partial crystallization. The resulting heterogeneous
mixture was stirred with heating at the same temperature
range for 1 h. The mixture was heated to reflux, and an
additional amount of 4 (13.88 g, 110 mmol) was added. After
the addition was complete, the stirred mixture was heated at
reflux for 5 h. The mixture was allowed to cool to 5 °C
with stirring and was kept at the same temperature for 0.5
h. The precipitated solid was collected by filtration under
suction, washed with EtOH (20 mL × 2), and air-dried with
heating at 50 °C to give 5 (42.37 g) as a white solid in 87.7%
yield. Mp 176.0-178.1 °C (lit.:5a mp 174-176 °C). IR ν-
(KBr) 3288, 3078, 3012, 2948, 2557, 2445, 2113, 1873,
1805, 1653, 1522, 1473, 1462, 1414, 1381, 1358, 1335, 1313,
1244, 1225, 1171, 1138, 1086, 1053, 1034, 1007, 976, 935,
water azeotropically until the residual volume diminished
to about 115 mL. The residue was filtered while it was kept
at 70 °C; the filtration suffered clogging unless it was
conducted at temperatures over 70 °C. The filter cake was
washed with warm PhMe (10 mL, 40 °C). The combined
filtrate and washing were concentrated in vacuo at 70 °C
(bath temperature) to give a pale-yellow and viscous oil
(about 34 g). Ethylcyclohexane (ECH; 67 mL) was added,
and the mixture was heated to 60-70 °C until the mixture
turned homogeneous. The homogeneous mixture was allowed
to cool to 5 °C over 1 h and kept at the same temperature
for 30 min. Precipitated solid was collected by filtration under
suction, washed with ice-chilled ECH (15 mL × 2), and air-
dried with heating at 45 °C for 3 h to give (()-6 (30.59 g)
in 90.1% yield as a white solid; prolonged heating should
be avoided since (()-6 showed a tendency to sublime. Mp
71.0-72.0 °C. IR ν (KBr) 3080, 2943, 2829, 1801, 1742,
1689, 1591, 1548, 1468, 1383, 1302, 1242, 1221, 1178, 1165,
1134, 1076, 1040, 1013, 924, 901, 847, 739, 690, 573, 544
cm-1. 1H NMR (CDCl3) δ 7.24 (br s, 1H), 7.21 (d, J ) 4.8
Hz, 1H), 6.98-6.80 (m, 2H), 5.19 (dd, J ) 4.0 Hz, 8.0 Hz,
1H), 2.70-2.50 (m, 2H), 2.29 (s, 6H), 2.00-1.88 (m, 2H).
(S)-N,N-Dimethyl-N-[3-hydroxy-3-(2-thienyl)propyl]-
ammonium (S)-Mandelate (8). To a stirred solution of
(()-6 (9.25 g, 50 mmol) in PhMe (92.5 mL) and MeOH
(2.3 mL) was added (S)-mandelic acid (7) (3.43 g, 23 mmol)
at room temperature. The resulting suspension was stirred
and heated to 80 °C to make the mixture homogeneous. The
stirring was continued with heating at 80 °C for 0.5 h and
the solution was allowed to cool to 25 °C at a rate of 0.5
°C/min. After stirring at 25 °C for 1 h, precipitated solid
was collected by filtration under suction. The filter cake was
washed once with PhMe (10 mL), air-dried with heating at
50 °C to give 8 (6.85 g) as a white solid in 41% yield. Mp
120.5-121.3 °C. [R]20D +30.4 (c 1.00, MeOH). Anal. Calcd
for C17H23NO4S: C, 60.51; H, 6.87; N, 4.15; S, 9.50.
Found: C, 60.3; H, 6.8; N, 4.1; S; 9.5. The enantiomeric
purity of (S)-6 contained in 8 was determined to be 93% ee
by HPLC as follows: To a portion (2 mg) of 8 was added
2 M aqueous solution of NaOH (0.2 mL) followed by MTBE
(0.2 mL). The layers were separated, and the MTBE layer
was concentrated in vacuo. To the residue was added i-PrOH
(2.0 mL), and a portion (1.0 µL) of the solution was injected
to a chromatograph running under the analytical condition
(2).
1
860, 852, 798, 768, 731, 654, 590, 500 cm-1. H NMR
(CDCl3) for the free amine δ 7.74 (d, J ) 3.6 Hz, 1H), 7.64
(d, J ) 4.8 Hz, 1H), 7.13 (t, J ) 4.0 Hz, 1H), 3.09 (t, J )
7.6 Hz, 2H), 2.76 (t, J ) 7.6 Hz, 2H), 2.29 (s, 6H).
(()-3-(N,N-Dimethylamino)-1-(2-thienyl)propan-1-ol (6).
To a suspension of NaBH4 (2.04 g, 54 mmol) in an aqueous
solution of NaOH (48% aqueous solution of NaOH, 16.50
g, 198 mmol; H2O, 118 mL) was added 5 (39.50 g, 180
mmol) in portions at room temperature over 1 h. The mixture
was stirred at room temperature for 3 h. NaBH4 (0.68 g, 18
mmol) was added, and the stirring was continued at room
temperature for 2 h. The mixture was stirred with heating at
69-71 °C for 1 h, during which the progress of the reaction
was monitored by HPLC as follows: From the reaction
mixture was taken an aliquot (30 µL), which was poured
into MTBE (0.3 mL). The layers were separated, and the
MTBE layer was concentrated in vacuo. To the residue was
added MeCN (2 mL), and a portion (1 µL) of the solution
was injected to a chromatograph running under the analytical
condition (1). On complete consumption of 5, the mixture
was cooled to 0 °C over 30 min. To the stirred mixture was
added 35% aqueous solution of HCl (30.97 g, 297 mmol)
carefully such that the inner temperature did not exceed 10
°C. The homogeneous mixture was washed with PhMe (10
mL × 2). To the aqueous layer was added 48% aqueous
solution of NaOH (about 12 g) with occasional cooling at
temperatures below 10 °C. The resulting turbid mixture was
seeded with a few crystals of (()-6 to induce crystallization.
Once crystals started to precipitate, 48% aqueous solution
of NaOH (29.71 g, 357 mmol in total; inclusive of the
amount added prior to the seeding) was added with cooling
such that the inner temperature did not exceed 25 °C. The
mixture was extracted with warm toluene (200 mL, 43-46
°C; 33 mL × 2, 30 °C). The combined toluene extracts (about
265 mL) were washed with saturated aqueous solution of
NaCl (5 mL × 2). The toluene solution was concentrated in
vacuo at 20 mmHg and 70 °C (bath temperature) to remove
Racemization of (R)-6. The filtrate (mother liquor)
obtained in the course of the above-mentioned diastereomeric
salt formation was analyzed by HPLC for the enantiomeric
purity of (R)-6, and it was determined to be 63% ee as
follows: From the filtrate was taken an aliquot (20 µL) to
which was added 2 M aqueous solution of NaOH (0.2 mL)
followed by PhMe (0.2 mL). The layers were separated, and
the PhMe solution was concentrated in vacuo. To the residue
was added i-PrOH (2.0 mL), and a portion (1.0 µL) of the
solution was injected to a chromatograph running under the
analytical condition (2). The filtrate [about 110 mL; a PhMe
solution of (R)-6 of 63% ee] was washed with 1 M aqueous
solution of NaOH (5 mL) and saturated aqueous solution of
Vol. 10, No. 5, 2006 / Organic Process Research & Development
•
911

NaCl (5 mL). The PhMe solution was concentrated in vacuo
until its volume diminished to 50 mL. To the residue was
added 1.5 M aqueous HCl solution (89 mL), and the resulting
two-phase mixture was stirred at 22 °C for about 7 h during
which the progress of the reaction was monitored by HPLC
as follows: From the aqueous layer was taken an aliquot
(20 µL) to which was added 2 M aqueous NaOH solution
(0.2 mL) followed by MTBE (0.2 mL). The layers were
separated, and the MTBE layer was concentrated in vacuo.
To the residue was added i-PrOH (2.0 mL), and a portion
(1.0 µL) of the solution was injected to a chromatograph
running under the analytical condition (2). When the enan-
tiomeric purity of (R)-6 diminished to below 5% ee, 48%
aqueous NaOH solution (13 mL) was added to the stirred
mixture. The layers were separated, and the PhMe layer was
washed with saturated aqueous solution of NaCl (5 mL). The
PhMe solution was concentrated in vacuo to give almost
racemized 6 as an off-white solid (5.30 g). Its enantiomeric
purity was determined to be 1.2% ee by HPLC as follows:
To an aliquot (1 mg) of the solid was added i-PrOH (2.0
mL), and a portion (1.0 µL) of the solution was injected to
a chromatograph running under the analytical condition (2).
It was also shown to be composed of 15 (5.4 area %), 6
(90.6 area %), and 14 (3.9 area %) by HPLC as follows:
To an aliquot (2 mg) of the solid was added MeCN (2.0
mL), and a portion (5.0 µL) of the solution was injected to
a chromatograph running under the analytical condition (3).
To the solid thus obtained (5.30 g) was added ECH (20 mL),
and the mixture was heated to 60 °C where it became
homogeneous. The resulting solution was allowed to cool
to 47 °C over 0.5 h. The solution was seeded with a crystal
of (()-6 and allowed to cool to 22 °C over 1 h. The
heterogeneous mixture was stirred at 22 °C for 1 h.
Precipitated solid was collected by filtration under suction,
washed with ECH (5 mL × 2), air-dried with heating at 45
°C to give (()-6 (4.60 g) in 86.8% yield as a white solid,
an overall yield from virgin (()-6 being 49.7%. The racemic
nature (0% ee) of the product was confirmed by HPLC
conducted under the analytical condition (2), and it was
shown to be composed of 15 (3.0 area %), 6 (97.0 area %),
and 14 (0 area %) by HPLC analysis conducted under the
analytical condition (3).
consumed in >99.5%. The mixture was allowed to cool to
room temperature, and ice-chilled H2O (25 mL) was added;
caution: when the aqueous mixture was left to stand at room
temperature for 16 h, (S)-18 produced suffered hydrolysis
to give back (S)-6 in 1-2% yield. The layers were separated,
and the PhMe layer was washed with saturated aqueous NaCl
solution (20 mL). The PhMe solution was concentrated in
vacuo at 50 mmHg and 60 °C (bath temperature) until its
volume reached 20 mL. To the residual PhMe solution were
added PhMe (20 mL) and solid NaHCO3 (0.25 g, 3.0 mmol).
The resulting heterogeneous mixture was stirred and heated
to reflux (bath temperature: 120 °C). A solution of EtOCOCl
(6.5 g, 60 mmol) in PhMe (12 mL) was added dropwise over
20 min with the refluxing PhMe being circulated through a
Dean-Stark apparatus. The mixture was stirred with heating
at reflux where the refluxing PhMe was kept to circulate
through the Dean-Stark apparatus. The progress of the
reaction was monitored by HPLC as follows: From the
reaction mixture was taken an aliquot (50 µL), which was
concentrated in vacuo. To the residue was added MeCN (2
mL), and a portion (2 µL) of the solution was injected to a
chromatograph running under the analytical condition (4).
When (S)-18 was consumed in >97%, the mixture was
allowed to cool to room temperature. H2O (20 mL) was
added, and the layers were separated. The PhMe layer was
washed with 0.5 M aqueous HCl solution (20 mL) and
saturated aqueous NaCl solution (20 mL × 2) and then was
concentrated in vacuo at 60 °C (bath temperature) to give
(S)-16 as a pale-yellow oil (8.50 g) in 90.1% yield. IR νmax
(KBr) 3491, 3107, 2982, 1747, 1699, 1487, 1404, 1371,
1254, 1005, 862, 841, 791, 771, 710 cm-1; 1H NMR (CDCl3)
δ 7.36-7.20 (m, 1H), 7.18-7.10 (m, 1H), 7.00-6.94 (m,
1H), 5.90-5.80 (m, 1H), 4.21-4.14 (m, 2H), 4.10 (q, J )
6.8 Hz, 2H) 3.50-3.20 (m, 2H), 2.89 (s, 3H), 2.38-2.10
(m, 2H), 1.29 (t, J ) 7.2 Hz, 3H), 1.24 (t, J ) 6.8 Hz, 3H).
MS m/z 338{[M + Na]+}. Chemical purity: 91.1% as
determined by HPLC under the analytical condition (4).
(S)-3-(N-Methylamino)-3-(2-thienyl)propan-1-ol (1). To
a stirred solution of (S)-16 (4.50 g, 14.3 mmol) in EtOH
(7.7 mL) and H2O (12.0 mL) was added 48% aqueous NaOH
solution (9.90 g, 11.9 mmol) dropwise. The mixture was
stirred and heated at 82 °C for 3 h. H2O (30 mL) was added,
and the mixture was extracted with PhMe (50 mL × 1, 30
mL × 1, 20 mL × 1) at 50 °C. The combined PhMe extracts
were warmed to 50 °C and washed with saturated aqueous
NaCl solution (20 mL). The PhMe solution was concentrated
in vacuo until its volume diminished to 50 mL. To the residue
was added Na2SO4 (0.3 g) to adsorb inorganic contaminants,
but not for drying. The mixture was filtrated, and the filtrate
was concentrated in vacuo to give crude (S)-1 as a pale-
yellow solid (2.60 g). ECH (11.25 mL) and PhMe (3.75 mL),
the ratio of ECH to PhMe being 3/1, were added, and the
mixture was heated to 60 °C where it became homogeneous.
The resulting solution was allowed to cool to 52 °C over
0.5 h. The solution was seeded with a crystal of (S)-1 and
allowed to cool to 25 °C over 1 h. The heterogeneous mixture
was stirred at 25 °C for 1 h. Precipitated solid was collected
by filtration under suction, washed with ECH (4 mL), and
(S)-3-(N-Ethoxycarbonyl-N-methyl)amino-1-ethoxy-
carbonyloxy-(2-thienyl)propane (16). To 8 [10.1 g, 30.0
mmol; containing (S)-6 of 93% ee] was added an aqueous
solution of NaOH (48% aqueous solution of NaOH, 3.00 g,
36.0 mmol; H2O, 12 mL). The mixture was extracted with
PhMe (50 mL). To the PhMe extract was added solid
NaHCO3 (5.04 g, 60 mmol). The heterogeneous mixture was
stirred and heated to 90 °C (inner temperature). A solution
of EtOCOCl (4.23 g, 39 mmol) in PhMe (12 mL) was added
dropwise over 15 min. The stirring was continued at 90 °C,
and the progress of the reaction was monitored by HPLC as
follows: From the reaction mixture was taken an aliquot
(50 µL), which was concentrated in vacuo. To the residue
was added MeCN (2 mL), and a portion (2 µL) of the
solution was injected to a chromatograph running under the
analytical condition (4). It took about 15 min until (S)-6 was
912
•
Vol. 10, No. 5, 2006 / Organic Process Research & Development

air-dried with heating at 45 °C to give (S)-1 (1.82 g) in 74.6%
portion (2.0 µL) of the 1,2-dicloroethane layer was injected
to a chromatograph running under the analytical condition
(6). When the presence of unconsumed (S)-1 or cyclic (S)-
N,O-acetal (19) was detected by the GLC analysis, a further
amount (100 mg) of NaBH(OAc)3 was added, and the stirring
was continued for another 1 h. To the mixture was added 3
M aqueous NaOH solution (4 mL). The mixture was
extracted with MTBE (5 mL). The MTBE extract was
concentrated in vacuo to give crude (S)-6 as a white solid in
about 90% yield. From the solid was taken an aliquot (1
mg) to which was added i-PrOH (2.0 mL). A portion (1.0
µL) of the solution was injected to a chromatograph running
under the analytical condition (2).
yield as a white solid. Mp 71.8-73.2 °C. [R]20 -13.3 (c
D
10.0, MeOH) [lit.:5c [R]22 -12.5 (c 4.4, MeOH)]. IR νmax
D
(KBr) 3288, 3106, 2947, 2893, 2839, 1493, 1472, 1433,
1358, 1306, 1274, 1221, 1180, 1110, 1088, 939, 908, 715,
1
695 cm-1. H NMR (CDCl3) δ 7.21 (dd, J ) 1.2 Hz, 5.2
Hz, 1H), 6.98-6.90 (m, 2H), 5.20 (dd, J ) 3.2 Hz, 8.4 Hz,
1H), 3.00-2.92 (m, 1H), 2.90-2.82 (m, 1H), 2.44 (s, 3H),
2.02-1.84 (m, 2H); signals due to two exchangeable protons
(NH and OH) were too broad to discern. Anal. Calcd for
C8H13NOS: C, 56.11; H, 7.65; N, 8.18; S, 18.72. Found:
C, 56.0; H, 7.6; N, 8.1; S, 18.7. The chemical purity of (S)-1
was determined to be 100% by HPLC as follows: To an
aliquot (2 mg) of (S)-1 was added MeCN (2.0 mL), and a
portion (5.0 µL) of the solution was injected to a chromato-
graph running under the analytical condition (5). The
enantiomeric purity of (S)-1 was determined to be 100% ee
as follows: To a stirred mixture of (S)-1 (34.3 mg, 0.2
mmol), NaBH(OAc)3 (212 mg, 1.0 mmol), and 1,2-chloro-
ethane (3 mL) was added 37% aqueous solution of formal-
dehyde (17.8 mg, 0.22 mmol) at room temperature. The
stirring was continued at room temperature for 1 h during
which the progress of the reaction was monitored by GLC
as follows: From the reaction mixture was taken an aliquot
(100 µL) which was poured into 1.0 M aqueous solution of
NaOH (0.5 mL). The mixture was agitated for 5 min, and a
(S)-3-Methyl-2-(2-thienyl)-perhydro-1,3-oxazine (19).
A solution of (S)-1 (10 mg) and paraformaldehyde (100 mg)
in MTBE was stirred at room temperature for 0.5 h. The
usual aqueous workup provided (S)-19. TLC (i-PrOH) Rf 0.07
1
for (S)-1 and 0.70 for (S)-19. H NMR (CDCl3) δ 7.27-
7.25 (m, 1H), 7.01-6.90 (m, 2H), 4.73 (dd, J ) 11.2 Hz,
2.4 Hz, 1H), 4.53 (dd, J ) 9.2 Hz, 2.0 Hz, 1H), 4.27 (d, J
) 11.2 Hz, 1H), 3.10-3.00 (m, 1H), 2.96s2.82 (m, 1H),
2.48 (s, 3H), 2.50-2.36 (m, 1H), 1.76-1.70 (m, 1H).
Received for review June 18, 2006.
OP060118L
Vol. 10, No. 5, 2006 / Organic Process Research & Development
•
913