Novel synthesis of butyl (S)-2-hydroxybutanoate, the key intermediate of PPARα agonist (R)-K-13675 from butyl (2 S,3 R)-epoxybutanoate and butyl (S)-2,3-epoxypropanoate

Article abstract of DOI:10.1055/s-0032-1318507

Novel synthetic methods for the production of butyl (S)-2-hydroxybutanoate starting from butyl (2S,3R)-epoxybutanoate or butyl (S)-2,3-epoxypropanoate were established. The former method utilized the regioselective thiolysis of the epoxybutanoate mediated by scandium triflate and subsequent reductive cleavage of the thioether to give butyl (S)-2-hydroxybutanoate with stereochemical retention in quantitative yield. The latter method utilized one-step conversion by a combination of methylmagnesium bromide and copper catalyst in high yield. Georg Thieme Verlag Stuttgart, New York.

Full text of DOI:10.1055/s-0032-1318507

PAPER
▌1319
paper
Novel Synthesis of Butyl (S)-2-Hydroxybutanoate, the Key Intermediate of
PPARα Agonist (R)-K-13675 from Butyl (2S,3R)-Epoxybutanoate and Butyl
(S)-2,3-Epoxypropanoate
Butyl (S)-2-Hydroxybutanoate
Takaaki Araki, Minoru Koura, Yukiyoshi Yamazaki, Kimiyuki Shibuya*
Tokyo New Drug Research Laboratories, Pharmaceutical Division, Kowa Co., Ltd., 2-17-43, Noguchicho, Higashimurayama, Tokyo 189-0022,
Japan
Fax +81(42)3950312; E-mail: k-sibuya@kowa.co.jp
Received: 08.02.2013; Accepted after revision: 06.03.2013
Although a number of synthetic methods4–7 have been re-
Abstract: Novel synthetic methods for the production of butyl (S)-
ported, most of them are noted as being technically diffi-
2-hydroxybutanoate starting from butyl (2S,3R)-epoxybutanoate or
cult requiring the preparation of special chiral ligands or
the tedious handling of enzymes and furthermore are cost
butyl (S)-2,3-epoxypropanoate were established. The former meth-
od utilized the regioselective thiolysis of the epoxybutanoate medi-
ated by scandium triflate and subsequent reductive cleavage of the prohibitive resulting in the production of waste products
thioether to give butyl (S)-2-hydroxybutanoate with stereochemical
retention in quantitative yield. The latter method utilized one-step
conversion by a combination of methylmagnesium bromide and
due to environmental regulations at pilot scale or larger.
In order to prevent potential racemization or avoid incom-
plete asymmetric induction, we decided to use chiral tem-
plates imbued with the requisite chiral center, such as
easily available butyl (2S,3R)-epoxybutanoate (5) and bu-
tyl (S)-2,3-epoxypropanoate (6)8 (Scheme 2, routes C and
D).
copper catalyst in high yield.
Key words: epoxides, regioselectivity, ring opening, Lewis acids,
scandium
Recently we developed the highly potent and selective
PPARα agonist, (R)-K-13675 (1).1 This agent has shown
promise in clinical trials as an excellent candidate for the
treatment of hyperlipidemia by reducing plasma triglycer-
ides and increasing HDL cholesterol. In this molecule, the
(aryloxy)butanoic acid is the key pharmacophore for the
potency and selectivity. Therefore, we regarded butyl (S)-
2-hydroxybutanoate (2) as a useful and key synthon in the
preparation of (R)-K-13675 (1) (Scheme 1).
O
H
O
O
On-Bu
OH
NH2
i) HNO2
i) RSH
ii) Raney Ni
5
ii) H+/n-BuOH
3
route C
route A
regioselective
ring-opening
O
On-Bu
H
OH
Our initial attempt prepared 2 from inexpensive (S)-2-
aminobutanoic acid (3),2a–d but this was accompanied par-
tial racemization (ca. 5% ee)2e (Scheme 2, route A). Next,
we attempted to establish a practical method to provide 2
via (S)-2-hydroxybutyrolactone (4) in a multi-kilogram-
production scale3 (Scheme 2, route B). However, the mul-
titude of reaction steps and high cost of goods in this
method creates limitations when scaling up to commercial
scale manufacture.
2
route D
O
route B
O
i) TMSI, n-BuOH
ii) H2, Pd
" Me "
O
On-Bu
O
H
OH
6
4
Scheme 2 Four synthetic approaches for production of butyl (S)-2-
hydroxybutanoate (2)
N
O
O
O
O
O
N
OH
On-Bu
MeO
H
OH
n-butyl (S)-2-hydroxybutanoate
PPARα agonist
(R)-K-13675 (1)
2
Scheme 1 Synthesis of (R)-K-13675 (1)
SYNTHESIS 2013, 45, 1319–1324
Advanced online publication: 03.04.2013
0
0
3
9
-
7
8
8
1
1
4
3
7
-
2
1
0
X
DOI: 10.1055/s-0032-1318507; Art ID: SS-2013-F0110-OP
© Georg Thieme Verlag Stuttgart · New York

1320
T. Araki et al.
PAPER
In the first approach, we envisaged the synthesis of 2 via proceeded by a transition state with the coordination of
the regioselective ring opening of butyl (2S,3R)-epoxybu- scandium triflate to both the epoxide and the ester carbon-
tanoate (5) using Lewis acids. Our plan was as follows: yl oxygen as shown in Scheme 3. When the bond is
The nucleophilic attack at the C3 position of 5 by areneth- formed between scandium triflate and the hydroxy group,
iols or alkanethiols and successive reductive cleavage of it was postulated that one equivalent of triflate was liber-
the resultant thioether to yield the desired 29,10 (Scheme 2, ated from scandium triflate. This is the reason that scandi-
route C). According to the literature,11a–d Riego and um triflate should be thought of as a mediator, not as the
Aggarwal both demonstrated the regioselective thiolysis recyclable catalyst as shown in Table 1 (entry 23). In the
of the 2,3-epoxy ester or amide in moderate yields.11a,c same way, due to the competitive coordination of Sc(III)
These methodologies were based on the coordination of with the oxygen of tetrahydrofuran, the thiolysis reaction
the Lewis acid to the carbonyl and epoxide oxygens and was inhibited when tetrahydrofuran was used as the sol-
then preferential nucleophilic attack at the C3 position. vent and no product resulted (entry 17).
The first reported method required two nucleophile equiv-
alents and excess Lewis acid, AlPO4-Al2O3. Also, these
TfOH
OTf
Sc(OTf)2
O
conditions result in a long reaction time because of the
low reactivity of AlPO4-Al2O3. The second reported
method resulted in the regioselective ring opening of
phenylglycidic amides by thiophenol in the presence of
ytterbium(III) triflate at –78 °C. Regrettably, the sub-
strates were restricted only to phenylglycidic amides that
can be easily attacked by the sulfur nucleophiles at the
more reactive benzylic position. The application of this
reaction scheme to other substrates is uncertain. These is-
sues prompted us to explore methods for 2,3-epoxy ester
ring opening with Lewis acids with high regioselectivity
and high yield under mild conditions.
Sc(OTf)2
H
O
O
H
O
S
H
H
R
On-Bu
On-Bu
SH
R
5
Scheme 3 Plausible transition state of 5 with scandium triflate
Next, we furthermore examined the use of other lantha-
nide triflates in the thiolysis reaction. Ytterbium(III) tri-
flate gave 7d in the same manner11a (entry 18). Although
Sharpless reported a satisfactory result when using lantha-
num(III) triflate with glycidic acids and secondary am-
ides,11b lanthanum(III) triflate and titanium(IV)
isopropoxide did not afford the desired product, 7d (entry
19 and 20). One speculation is that the coordination of ti-
tanium(IV) isopropoxide with the ester might be different
from that for acids or amides. In spite of further screening
of additional Lewis acids, no other reagents were found to
be suitable for the desired reaction.
A variety of Lewis acids were screened in the reaction of
5 with ethanethiol (Table 1). Starting with typical Lewis
acids containing chloride atoms (entry 1–10), some reac-
tions gave the chloride adduct 9 rather than the thioether
7a or 8a in dichloromethane (entries 1, 3, 5, and 9). This
observation contrasts with previous reports11e–h that the
use of chloro-substituted Lewis acids (InCl3, AlCl3,
SnCl2, SnCl4, ZnCl2, MgCl2, and CaCl2) successfully
yielded the thiol adduct and not the chloro adduct. The re-
actions in tetrahydrofuran (entries 4, 6, 8, and 10) resulted
in either complex mixtures or in no reaction product ex-
cept in the case of zinc chloride12 (entry 2). We postulate
that the nucleophilic thiol13 is associated with strong Lew-
is acids and the resulting chloride ion liberated subse-
quently attacked the C3 position of the epoxy ester. In
order to avoid the formation of the chloride adduct, a
Lewis acid composed of a less nucleophilic ligand such as
triflate was selected for further investigation.
Based upon the data generated in this series of experi-
ments, we concluded that scandium triflate is the most ef-
fective Lewis acid for regioselective thiolysis of 5 under
mild conditions. Thus the obtained butyl (2R,3S)-3-(aryl-
alkylthio)-2-hydroxybutanoates 7b–d were successively
reduced by Raney nickel in the presence of Celite to af-
ford the desired 2 with complete retention of optical
purity17 (Scheme 4).
Stoichiometric amounts of scandium triflate14 were found
to be very effective and the regioselective ring opening re-
action resulted to give 7a in satisfactory yield (entry 13).
The combinations with other thiols such as dodecanethi-
ol,15 benzyl mercaptan, and thiophenol also afforded the
desired products 7b–d in the same manner (entries 14–
16). Thiophenol with stronger acidity compared to other
alkanethiols gave a better yield, but it had the potential
disadvantage that the successive desulfuration generates
benzene as an unwanted byproduct.16
RSH (1 equiv)
Sc(OTf)3 (1 equiv)
SR
O
O
O
H
CH2Cl2, r.t.
83–95%
On-Bu
H
On-Bu
OH
5
7b R = C12H25
7c R = Bn
7d R = Ph
(98.1% ee)
Raney Ni
Celite®
O
On-Bu
EtOH, 80 °C
78–100%
OH
At first glance, scandium triflate can be expected to serve
a catalytic function. However, these results suggested that
the regioselective thiolysis at C3 of 2,3-epoxybutanoate
2
(98.1% ee)
Scheme 4 Synthesis of butyl (S)-2-hydroxybutanoate (2)
Synthesis 2013, 45, 1319–1324
© Georg Thieme Verlag Stuttgart · New York

PAPER
Butyl (S)-2-Hydroxybutanoate
1321
A further study demonstrated that the scandium triflate Considering the above issue, our attention was directed to
can be easily recovered after workup, and that it can be re- catalytic reactions and one-step conversions. First, we re-
cycled and reused without losing activity several times.18 acted 6 with low-order lithium dimethylcuprate and suc-
Thus this result would be environmentally favorable to ceeded in obtaining 2 with high regioselectivity in
minimize industrial waste. Herewith we successfully es- excellent yield (Table 2, entry 1). Next, we attempted to
tablished a useful method for the production of 2, except minimize the requisite amount of copper consumption by
that it requires an equivalent amount of scandium triflate investigating several copper catalysts including, copper(I)
even though it is a relatively inexpensive reagent.
iodide, copper(I) chloride, and copper(I) bromide in com-
bination with a methylmagnesium bromide.19 Of the cop-
Table 1 Ring-Opening Reaction of Epoxide 5 with Thiols and Lewis Acids
SR
O
OH
O
Cl
O
O
RSH
O
Lewis acid
H
On-Bu
On-Bu
On-Bu
On-Bu
solvent, r.t.
H
OH
7a–d
SR
8a d
OH
5
a
9
R = Et
–
b R = C12H25
c R = Bn
d R = Ph
Entry
RSH (equiv)
Lewis acid (equiv)
Solvent
Time (h)
Yield (%)
7/8/9
Recovery of 5b
1
2
Et (1.2)
Et (1.2)
Et (1.2)
Et (1.2)
Et (1.2)
Et (1.2)
Et (1.2)
Et (1.2)
Et (1.2)
Et (1.2)
Ph (1.0)
Ph (1.0)
Et (1.0)
C12H25 (1.0)
Bn (1.0)
Ph (1.0)
Ph (1.0)
Ph (1.0)
Ph (1.0)
Ph (1.0)
Ph (1.0)
Ph (1.0)
Ph (1.0)
ZnCl2 (1.2)
CH2Cl2
THF
20
20
20
20
3
trace:0:98a
0:0:46a
0
50
0
ZnCl2 (1.2)
3
ZrCl4 (1.2)
CH2Cl2
THF
trace:0:93a
complex mixture
17:0:83a
4
ZrCl4 (1.2)
–
5
AlCl3 (1.2)
CH2Cl2
THF
0
6
AlCl3 (1.2)
20
3
complex mixture
complex mixture
complex mixture
0:0:40a
–
7
Et2AlCl (1.2)
Et2AlCl (1.2)
VCl3 (1.2)
CH2Cl2
THF
–
8
20
20
20
24
12
7
–
9
CH2Cl2
THF
54
–
10
11
12
13
14
15
16
17
18
19
20
21
22
23
VCl3 (1.2)
no reaction
no reaction
complex mixture
83:0:0b
Zn(OTf)2 (1.0)
Al(OTf)3 (1.0)
Sc(OTf)3 (1.0)
Sc(OTf)3 (1.0)
Sc(OTf)3 (1.0)
Sc(OTf)3 (1.0)
Sc(OTf)3 (1.0)
Yb(OTf)3 (1.0)
La(OTf)3 (1.0)
Ti(Oi-Pr)4 (1.0)
BF3–Et2O (1.0)
Al2O3 (1.0)
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
THF
–
–
0
10
10
10
24
24
24
24
8
85:0:0b
0
90:0:0b
0
95:0:0b
0
no product
81:0:0b
0
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
0
no reaction
no reaction
34:trace:0b
trace:19:0b
26:trace:0
–
–
–
24
24
0
Sc(OTf)3 (0.2)
70
a Yield was determined by 1H NMR analysis.
b Isolated yield.
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2013, 45, 1319–1324

1322
T. Araki et al.
PAPER
AB SCIEX API4000 QTRAP apparatus. Elemental analyses (C, H,
N) were performed by Yanaco MT-5. The enantiopurities were de-
termined by HPLC analysis (Shimadzu LC-10Avp equipment,
CHIRALPAK® AD column, Daicel Chemical Industries, Ltd.).
Table 2 Reaction of Butyl (S)-2,3-Epoxypropanoate (6) with Meth-
ylmagnesium Bromide in the Presence of a Copper Catalyst
O
O
MeMgBr (1.5 equiv)
O
[Cu] cat.
On-Bu
On-Bu
Butyl (2R,3S)-3-(Ethylthio)-2-hydroxybutanoate (7a); Typical
Procedure
Et2O, –78 °C
OH
H
To a stirred soln of butyl (2S,3R)-2,3-epoxybutanoate (98.1% ee)
(5, 100 mg, 0.632 mmol) and Sc(OTf)3 (311 mg, 0.632 mmol) in
CH2Cl2 (5 mL) was added EtSH (39.3 mg, 0.632 mmol) at r.t. The
mixture was stirred for 7 h and then diluted with H2O (20 mL) and
extracted with CHCl3 (3 × 20 mL). The combined organic layers
were dried (Na2SO4), and concentrated in vacuo. The residue was
purified by column chromatography (silica gel, n-hexane–EtOAc,
5:1) to give 7a (116 mg, 83%) as a colorless oil; [α]D +10.0 (c
0.28, CHCl3).
IR (ATR, neat): 3490, 2962, 2873, 1732, 1456, 1209, 1130 cm–1.
1H NMR (400 MHz, CDCl3): δ = 0.95 (t, J = 7.6 Hz, 3 H), 1.24 (d,
J = 6.8 Hz, 3 H), 1.28 (t, J = 7.6 Hz, 3 H), 1.40 (sextet, J = 7.6 Hz,
2 H), 1.63–1.70 (m, 2 H), 2.57–2.71 (m, 2 H), 3.00 (br ds, 1 H), 3.20
(qd, J = 6.8, 2.8 Hz, 1 H), 4.17–4.27 (m, 2 H), 4.33 (d, J = 2.8 Hz,
1 H).
13C NMR (100 MHz, CDCl3): δ = 13.6, 14.7, 15.7, 19.0, 25.2, 30.5,
43.4, 65.7, 73.2, 172.8.
6
2
Entry
Cu source
Cu (equiv)
Time
Yield (%)
1
2
3
4
5
6
7
8
LiCuMe2
CuI
2.0
15 min
3.5 h
99
83
99
79
76
–
0.075
0.15
0.15
0.15
0
25
CuI
15 min
15 min
15 min
2.5 h
CuCl
CuBr
CuI
CuI
0.75
1.5
15 min
2.5 h
41
–
CuI
HRMS (EI): m/z [M]+ calcd for C10H20O3S: 220.1133; found:
220.1141.
per catalysts, copper(I) iodide was found to produce the
highest yield as shown in Table 2 (entries 3–5). Interest-
ingly, the chemical yield of 2 was not linear, doubling the
amount of copper(I) iodide (entries 7 and 8) did not dou-
ble the yield. The desired product 2 was obtained with
high regioselectivity in quantitative yield when 0.15
equivalents of copper(I) iodide were used (entry 3).
Anal. Calcd for C10H20O3S: C, 54.51; H, 9.15. Found: C, 54.45; H,
9.18.
Butyl (2R,3S)-3-(Dodecylthio)-2-hydroxybutanoate (7b)
Following the typical procedure using 5 (50 mg, 0.316 mmol) and
Sc(OTf)3 (156 mg, 0.316 mmol) in CH2Cl2 (4 mL) and dodecane-1-
thiol (64.0 mg, 0.316 mmol); time: 10 h. Work up and purification
by column chromatography (silica gel, n-hexane–EtOAc, 5:1) gave
7b (97.0 mg, 85%) as a colorless oil; [α]D25 +9.34 (c 0.77, CHCl3).
By using a catalytic amount of copper(I) iodide and meth-
ylmagnesium bromide, we achieved one-step conversion
of 6 into 2 in excellent yield without affecting the ester.20
IR (ATR, neat): 3471, 2924, 2854, 1733, 1458, 1210, 1129 cm–1.
1H NMR (400 MHz, CDCl3): δ = 0.88 (t, J = 6.9 Hz, 3 H), 0.95 (t,
J = 7.6 Hz, 3 H), 1.23–1.45 (m, 23 H), 1.59 (quintet, J = 7.6 Hz, 2
H), 1.67 (quintet, J = 6.9 Hz, 2 H), 2.54–2.66 (m, 2 H), 2.99 (br ds,
1 H), 3.17 (dq, J = 6.9, 3.2 Hz, 1 H), 4.16–4.27 (m, 2 H), 4.32 (d,
J = 2.8 Hz, 1 H).
13C NMR (100 MHz, CDCl3): δ = 13.6, 14.0, 15.7, 19.0, 22.6, 28.9,
29.2, 29.3, 29.5, 29.5, 29.6, 29.6, 29.6, 30.5, 31.3, 31.9, 43.8, 65.6,
73.2, 172.8.
In summary, we have established two methods for the
synthesis of butyl (S)-2-hydroxybutanoate (2) as a key in-
termediate of (R)-K-13675 (1) starting from butyl
(2S,3R)-epoxybutanoate (5) and butyl (S)-2,3-epoxypro-
panoate (6). The method from butyl (2S,3R)-epoxybu-
tanoate (5) is a useful method of production, unfortunately
the method requires large amounts of scandium triflate al-
beit that it is inexpensive and recyclable. The method
from butyl (S)-2,3-epoxypropanoate uses a Grignard
method with catalytic copper reagent and provides an ef-
ficient conversion to give 2 in excellent yield. Both results
provide a useful demonstration of the nucleophilic open-
ing of 2,3-substituted epoxy esters in a highly regioselec-
tive manner. Furthermore, we are, at present, refining a
more practical and environmentally friendly production.
HRMS (EI): m/z [M]+ calcd for C20H40O3S: 360.2698; found:
360.2684.
Anal. Calcd for C20H40O3S: C, 66.62; H, 11.18; S, 8.89. Found: C,
66.66; H, 11.35; S, 8.83;
Butyl (2R,3S)-3-(Benzylthio)-2-hydroxybutanoate (7c)
Following the typical procedure using 5 (300 mg, 1.90 mmol) and
Sc(OTf)3 (933 mg, 1.90 mmol) in CH2Cl2 (10 mL) and BnSH (234
mg, 1.90 mmol); time: 10 h. Workup used H2O (10 mL) and CHCl3
(2 × 10 mL). Purification by column chromatography (silica gel, n-
hexane–EtOAc, 5:1) gave 7c (484 mg, 90%) as a colorless oil;
[α]D20 –9.4 (c 1.5, CHCl3).
Commercially available reagents and solvents were used without
further purification. All reactions were carried out under an argon
or a nitrogen atmosphere unless otherwise noted. TLC analyses
were carried out on silica gel 60 F254 plates (Merck). Optical rota-
tions were measured on a JASCO P-2200 polarimeter with sodium
(D line) lamp. IR spectra were recorded on a Thermo Nicolet 370
IR (ATR, neat): 3470, 2961, 1732, 1454, 1209, 1127, 700 cm–1.
1H NMR (400 MHz, CDCl3): δ = 0.94 (t, J = 7.3 Hz, 3 H), 1.18 (d,
J = 7.3 Hz, 3 H), 1.36 (sextet, J = 7.3 Hz, 2 H), 1.61–1.67 (m, 2 H),
2.95 (br ds, 1 H), 3.09 (dq, J = 2.8, 7.3 Hz, 1 H), 3.79 (d, J = 13.5
Hz, 1 H), 3.85 (d, J = 13.5 Hz, 1 H), 4.14–4.24 (m, 2 H), 4.30 (d,
J = 2.7 Hz, 1 H), 7.22–7.36 (m, 5 H).
13C NMR (100 MHz, CDCl3): δ = 13.5, 15.4, 19.0, 30.4, 35.7, 42.9,
65.6, 73.1, 127.0 (2 ×), 128.4, 128.8 (2 ×), 137.9, 172.8.
1
FT-IR (ATR) spectrophotometer. H NMR spectra and 13C NMR
spectra were obtained on a JEOL JNM-AL 400 MHz spectrometer
using CDCl3 as solvent with tetramethylsilane as internal standard
unless otherwise noted. Mass spectra were obtained on a JEOL GC-
mate MS-BU20, JEOL JMS-T100GCV, Waters Xevo G2 QTof,
Synthesis 2013, 45, 1319–1324
© Georg Thieme Verlag Stuttgart · New York

PAPER
Butyl (S)-2-Hydroxybutanoate
1323
HRMS (EI): m/z [M]+ calcd for C15H22O3S: 282.1290; found:
282.1291;
1H NMR (400 MHz, CDCl3): δ = 0.90 (t, J = 7.3 Hz, 3 H), 1.32 (sex-
tet, J = 7.3 Hz, 2 H), 1.40 (d, J = 6.4 Hz, 3 H), 1.52–1.60 (m, 2 H),
2.67 (br ds, 1 H), 3.59 (d, J = 7.3 Hz, 1 H), 4.09–4.15 (m, 3 H),
7.29–7.34 (m, 3 H), 7.46–7.50 (m, 2 H).
13C NMR (100 MHz, CDCl3): δ = 13.5, 18.8, 20.5, 30.3, 57.5, 65.0,
67.7, 127.9 (2 ×), 128.9 (2 ×), 132.6, 133.0, 171.6.
Butyl (2R,3S)-2-Hydroxy-3-(phenylthio)butanoate (7d)
Following the typical procedure using 5 (100 mg, 0.632 mmol) and
Sc(OTf)3 (311 mg, 0.632 mmol) in CH2Cl2 (5 mL) and PhSH (69.6
mg, 0.632 mmol); time 10 h. Workup and purification by column
chromatography (silica gel, n-hexane–EtOAc, 5:1) gave 7d (117
mg, 95%) as a colorless oil; [α]D25 +42.8 (c 0.95, CHCl3).
HRMS (EI): m/z [M]+ calcd for C14H20O3S: 268.1133; found:
268.1133.
IR (ATR, neat): 3484, 2961, 2873, 1732, 1439, 1211, 1129 cm–1.
Chloride Adduct 9 (Mixture of Diastereomers) (Table 1, Entries
1–3, 5, and 9)
1H NMR (400 MHz, CDCl3): δ = 0.94 (t, J = 7.6 Hz, 3 H), 1.27 (d,
J = 6.8 Hz, 3 H), 1.38 (sextet, J = 7.6 Hz, 2 H), 1.65 (quintet, J = 7.6
Hz, 2 H), 2.98 (d, J = 4.6 Hz, 1 H), 3.65 (dq, J = 6.8, 2.8 Hz, 1 H),
4.14–4.25 (m, 2 H), 4.28 (br ds, 1 H), 7.24–7.34 (m, 3 H), 7.48 (dt,
J = 1.4, 6.6 Hz, 2 H).
IR (ATR, neat): 3460, 2962, 1737, 1219 cm–1.
1H NMR (400 MHz, CDCl3): δ = 0.95 (t, J = 7.6 Hz, 2.6 H), 1.24 (d,
J = 7.3 Hz, 0.4 H), 1.28 (t, J = 7.3 Hz, 0.4 H), 1.40 (sextet, J = 7.3
Hz, 2 H), 1.54 (d, J = 6.8 Hz, 2.6 H), 1.64–1.71 (m, 2 H), 2.59–2.69
(m, 0.3 H), 3.13–3.23 (m, 0.7 H), 4.18–4.35 (m, 4 H).
13C NMR (100 MHz, CDCl3): δ = 13.6, 14.8, 19.0, 30.5, 47.0, 65.9,
13C NMR (100 MHz, CDCl3): δ = 13.5, 14.6, 15.6, 19.0, 19.6, 25.2,
30.4, 30.5, 43.4, 58.5, 65.6, 66.1, 73.2, 74.7, 171.3, 172.8.
72.3, 127.4, 129.1 (2 ×), 132.2 (2 ×), 134.1, 172.8.
MS (EI): m/z 268 [M]+.
HRMS (EI): m/z [M + H]+ calcd for C8H16ClO3: 195.0788; found:
195.0777.
Anal. Calcd for C14H20O3S: C, 62.66; H, 7.51; S, 11.95. Found: C,
62.66; H, 7.61; S, 11.70.
Butyl (2R,3S)-3-(Benzylthio)-2-hydroxybutanoate (7c); Scaled-
Up Synthesis
Butyl (S)-2-Hydroxybutanoate (2)
To a stirred soln of butyl (2R,3S)-2-hydroxy-3-(phenylthio)butano-
ate (7d, 60.0 mg, 0.224 mmol) in EtOH (2 mL) was added Celite
(240 mg) and a suspension of Raney Ni (360 mg) in EtOH (3 mL)
at r.t. The mixture was stirred at 80 °C for 10 h and then filtered
through a pad of Celite. The filtrate was concentrated in vacuo, di-
luted with H2O (20 mL) and extracted with Et2O (3 × 20 mL). The
combined organic layers were dried (MgSO4) and concentrated in
vacuo. The residue was purified by column chromatography (silica
gel, CH2Cl2) to give 2 (35.8 mg, 100%) as a colorless oil; 98.1% ee.
To a stirred soln of butyl (2S,3R)-2,3-epoxybutanoate (98.1% ee)
(5, 80.0 g, 506 mmol) and Sc(OTf)3 (249 g, 506 mmol) in CH2Cl2
(1 L) was added BnSH (62.8 g, 506 mmoL) at r.t. The mixture was
stirred for 10 h. The mixture was diluted with H2O (1 L) and extract-
ed with CHCl3 (2 × 500 mL). The combined organic layers were
dried (Na2SO4, 100 g) and concentrated in vacuo. The residue was
purified by column chromatography (silica gel, n-hexane–EtOAc,
5:1) to give 7c (105 g, 74%) as a colorless oil. Spectral data was
identical with butyl (2R,3S)-3-(benzylthio)-2-hydroxybutanoate
(7c).
Butyl (2R,3S)-3-(dodecylthio)-2-hydroxybutanoate (7b, 400 mg,
1.11 mmol) or butyl (2R,3S)-3-(benzylthio)-2-hydroxybutanoate
(7c, 300 mg, 1.06 mmol) were used instead of 7d in this procedure
giving 2 (139 mg, 78% from 7b; 159 mg, 94%, from 7c).
Butyl (S)-2-Hydroxybutanoate (2); Scaled-Up Synthesis
To a stirred soln of butyl (2R,3S)-3-(benzylthio)-2-hydroxybutano-
ate (7c, 5.00 g, 17.7 mmol) in EtOH (150 mL) was added Celite
(20.0 g) and a suspension of Raney Ni (30 g) in EtOH (30 mL) at
r.t. The mixture was stirred at 80 °C for 10 h and then filtered
through a pad of Celite (20 g). The filtrate was concentrated in vac-
uo. The residue was purified by column chromatography (silica gel,
CH2Cl2) to give 2 (2.98 g, 100%) as a colorless oil; 98.1% ee. Spec-
tral data was identical with butyl (S)-2-hydroxybutanoate (2).
Optical purity of 2 was determined by HPLC as its 4-nitrobenzoate
according to the reported method.3b
HPLC (Daicel Chiralpak AD, column temperature; 35 °C; hexane–
EtOH, 60:40; flow rate; 1.0 mL/min): tR = 4.85 (R-isomer), 6.55
min (S-isomer).
[α]D20 –3.53 (c 1.01, CHCl3) [Lit.21 [α]D25 –3.4 (c 1.0, CHCl3)].
IR (ATR, neat): 3468, 2962, 1731, 1463, 1244, 1206, 1132 cm–1.
Demonstration for the Thiolysis Reaction of 5 with Recycled
Scandium Triflate
After the reaction and workup, the consumed Sc(OTf)3 could be re-
covered quantitatively as follows: The aqueous layer was concen-
trated and then dried in vacuo at 140 °C for 2 h.
1H NMR (400 MHz, CDCl3): δ = 0.95 (t, J = 7.6 Hz, 3 H), 0.97 (t,
J = 7.6 Hz, 3 H), 1.39 (qt, J = 7.6, 7.3 Hz, 2 H), 1.61–1.74 (m, 3 H),
1.79–1.89 (m, 1 H), 2.81 (d, J = 5.6 Hz, 1 H), 4.13–4.25 (m, 3 H).
13C NMR (100 MHz, CDCl3): δ = 9.0, 13.7, 19.1, 27.6, 30.6, 65.2,
To a stirred soln of butyl (2S,3R)-2,3-epoxybutanoate (5, 500 mg,
3.16 mmol) and thus recovered Sc(OTf)3 (1.56 g, 3.16 mmol) in
CH2Cl2 (30 mL) was added PhSH (348 mg, 3.16 mmol) at r.t. The
mixture was stirred for 4 h and then diluted with H2O (30 mL) and
extracted with CHCl3 (3 × 20 mL). The combined organic layers
were dried (Na2SO4) and concentrated in vacuo. The residue was
purified by column chromatography (silica gel, n-hexane–EtOAc,
5:1) to give 7d (807 mg, 95%) as a colorless oil; 98.1% ee.
71.5, 175.4.
HRMS (FAB): m/z [M + H]+ calcd for C8H17O3: 161.1178; found:
161.1181.
Anal. Calcd for C8H16O3: C, 59.97; H, 10.07. Found: C, 59.70; H,
9.80.
Butyl (2R,3R)-3-Hydroxy-2-(phenylthio)butanoate (8d)
To a stirred soln of butyl (2S,3R)-2,3-epoxybutanoate (5, 100 mg,
0.632 mmol) and alumina (64.4 mg, 0.632 mmol) in CH2Cl2 (5 mL)
was added PhSH (69.6 mg, 0.632 mmol) at r.t. The mixture was
stirred for 24 h and then diluted with H2O (20 mL) and extracted
with CHCl3 (3 × 20 mL). The combined organic layers were dried
(Na2SO4) and concentrated in vacuo. The residue was purified by
column chromatography (silica gel, n-hexane–EtOAc, 5:1) to give
8d (32.2 mg, 19%) as a colorless oil.
Butyl (S)-2-Hydroxybutanoate (2)
To a stirred soln of CuI (594 mg, 3.12 mmol) in anhyd Et2O (20 mL)
was added 0.52 M MeMgBr in THF (60.2 mL, 31.3 mmol) at
–15 °C and the mixture was stirred for 20 min under an argon atmo-
sphere. To the mixture was added a soln of butyl (S)-2,3-epoxypro-
panoate (97.0% ee) (6, 3.00 g, 20.8 mmol) in Et2O (8 mL) at –78 °C.
The mixture was stirred at this temperature for 20 min and then di-
luted with sat. aq NH4Cl (50 mL) and extracted with Et2O (2 × 50
mL). The combined organic layers were washed with brine (40
mL), dried (Na2SO4), and concentrated in vacuo. The residue was
[α]D20 +147.0 (c 1.1, CHCl3).
IR (ATR, neat): 3466, 2961, 1729, 1160, 691 cm–1.
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2013, 45, 1319–1324

1324
T. Araki et al.
PAPER
purified by column chromatography (silica gel, n-hexane–EtOAc,
5:1) to give 2 (3.37 g, 99%) as a colorless oil; 97.0% ee.
G.; Hall, R. F.; Perchonock, C. D.; Erhard, K. F.; Frazee, J.
S.; Ku, T. W.; Kondrad, K.; McCarthy, M. E.; Mong, S.
J. Med. Chem. 1987, 30, 959.
(10) Refer to the following articles: (a) Koura, M.; Yamazaki, Y.;
Shibuya, K. WO 2007013555, 2007. (b) Compagnone, R.
S.; Rapoport, H. J. Org. Chem. 1986, 51, 1713.
Acknowledgment
We thank Dr. S. Tanabe, Director of Tokyo New Drug Research La-
boratories, Kowa Co., Ltd, for his support and also Dr. Kevin Swiss
for his help in preparing the manuscript.
(c) Romanelli, M. N.; Bartolini, A.; Bertucci, C.; Dei, S.;
Ghelardini, C.; Giovannini, M. G.; Gualtieri, F.; Pepeu, G.;
Scapecchi, S.; Teodori, E. Chirality 1996, 8, 225.
(11) (a) Riego, J.; Costa, A.; Saa, J. M. Chem. Lett. 1986, 1565.
(b) Chong, J. M.; Sharpless, K. B. J. Org. Chem. 1985, 50,
1560. (c) Aggarwal, V. K.; Charmant, J. P. H.; Fuentes, D.;
Harvey, J. N.; Hynd, G.; Ohara, D.; Picoul, W.; Robiette, R.;
Smith, C.; Vasse, J. L.; Winn, C. L. J. Am. Chem. Soc. 2006,
128, 2105. (d) Flisak, J. R.; Gombatz, K. J.; Holmes, M. M.;
Jarmas, A. A.; Lantos, I.; Mendelson, W. L.; Novack, V. J.;
Remich, J. J.; Snyder, L. J. Org. Chem. 1993, 58, 6247.
(e) Puschmann, I.; Erker, T. Heterocycles 1995, 41, 709.
(f) Hashiyama, T.; Inoue, H.; Takeda, M.; Murata, S.;
Nagao, T. Chem. Pharm. Bull. 1985, 33, 2348.
References
(1) (a) Yamazaki, Y.; Abe, K.; Toma, T.; Nishikawa, M.;
Ozawa, H.; Okuda, A.; Araki, T.; Oda, S.; Inoue, K.;
Shibuya, K.; Staels, B.; Fruchart, J. C. Bioorg. Med. Chem.
Lett. 2007, 17, 4689. (b) Kitajima, K.; Miura, S.; Matsuo, Y.;
Uehara, Y.; Saku, K. Atherosclerosis 2009, 203, 75.
(2) (a) Jares-Erijiman, E. A.; Ingrum, A. L.; Carney, J. R.;
Rinehart, K. L.; Sakai, R. J. Org. Chem. 1993, 58, 4805. This
mechanism presumably involves anchimeric assistance of
the carboxy group in the decomposition of an intermediate
diazonium ion, leading to an α-hydroxy acid with retention
of configuration. See the following reference: (b) Brewster,
F.; Hiron, F.; Hughes, E. D.; Ingold, C. K.; Rao, P. A. D. S.
Nature (London) 1950, 166, 179. (c) Koga, K.; Taniguchi,
M.; Yamada, S. Tetrahedron Lett. 1971, 12, 263.
(d) Gringore, O. H.; Rouessac, F. P. Org. Synth. 1985, 63,
121. The reaction of butyl (S)-2-bromobutanoate obtained
via diazotization of (S)-2-aminobutanoic acid with phenol
accompanied with racemization (~81% ee): (e) Sugita, M.
JP 2011093869, 2011.
(g) Hashiyama, T.; Inoue, H.; Konda, M.; Takeda, M.
J. Chem. Soc., Perkin Trans. 1 1984, 1725. (h) Ranu, B. C.;
Mandal, T. Can. J. Chem. 2006, 84, 762.
(12) We postulated that THF can be cleaved under these
conditions. Whereas ZnCl2 is too weak a Lewis acid to open
the THF ring.
(13) March, J. Advanced Organic Chemistry, 4th ed.; Wiley-
VCH: Weinheim, 1992, 350.
(14) Sc(OTf)3 in bulk for multi-kg scale is commercially
available from Shanghai Synchem Technology and
Sinofluoro Scientific Co., Ltd. (China).
(3) (a) Yamazaki, Y.; Araki, T.; Koura, M.; Shibuya, K.
Synthesis 2008, 1017. (b) Yamazaki, Y.; Araki, T.; Koura,
M.; Shibuya, K. Tetrahedron 2008, 64, 8155.
(4) Davis, F. A.; Chen, B. C. Chem. Rev. 1992, 92, 919.
(5) (a) Kirschbaum, B.; Fell, R. WO 03005010, 2003.
(b) Senba, N. JP 2000217590, 2000.
(6) (a) Nakamura, K.; Inoue, K.; Ushio, K.; Oka, S.; Ohno, A.
J. Org. Chem. 1988, 53, 2589. (b) Nakamura, K.;
Yamanaka, R.; Matsuda, T.; Harada, T. Tetrahedron:
Asymmetry 2003, 14, 2659.
(7) Burk, M. J.; Kalberg, C. S.; Pizzano, A. J. Am. Chem. Soc.
1998, 120, 4345.
(8) Both compounds are commercially available from Osaka
Organic Chemical Industry Ltd.; they can be easily prepared
from the corresponding racemic epoxypropanoate and trans-
2,3-epoxybutanoate by the treatment with protease: Akaishi,
R.; Tsutsumi, A.; Araki, S.; Hirose, Y.; Ohno, A. JP
2000166592, 2000.
(9) (a) Behrens, C. H.; Sharpless, K. B. J. Org. Chem. 1985, 50,
5696. (b) Chong, J. M.; Sharpless, K. B. Tetrahedron Lett.
1985, 26, 4683. (c) Amantini, D.; Fringuelli, F.; Pizzo, F.;
Vaccaro, L. J. Org. Chem. 2001, 66, 4463. (d) Gleason, J.
(15) Node, M.; Kumar, K.; Nishide, K.; Ohsugi, S.; Miyamoto, T.
Tetrahedron Lett. 2001, 42, 9207.
(16) Benzene in pharmaceutical products must be controlled to
less than 2 ppm as organic volatile impurities under the ICH
Q3C guidelines.
(17) The optical purity of 2 was determined by converting it into
its 4-nitrobenzoate and then examining it by HPLC analysis,
see also ref. 3b.
(18) Refer to the following articles: (a) Kobayashi, S. Chem. Lett.
1991, 2187. (b) Barrett, A. G. M.; Braddock, D. C. Chem.
Commun. 1997, 351.
(19) (a) Ashby, E. C.; Smith, R. S.; Goel, A. B. J. Org. Chem.
1981, 46, 5133. (b) Larcheveque, M.; Perit, Y. Tetrahedron
Lett. 1987, 28, 1993.
(20) (a) Perrone, M. G.; Santandrea, E.; Nunno, L. D.; Scilimati,
A.; Tortorella, V.; Capitelli, F.; Bertolasi, V. Tetrahedron:
Asymmetry 2005, 16, 783. (b) Diaz, M.; Ibarzo, J.; Ortuno,
R. M. Tetrahedron: Asymmetry 1994, 5, 37. (c) Diaz, M.;
Ortuno, R. M. Tetrahedron: Asymmetry 1995, 6, 1845.
(21) Parida, S.; Dordick, J. S. J. Am. Chem. Soc. 1991, 113, 2253.
Synthesis 2013, 45, 1319–1324
© Georg Thieme Verlag Stuttgart · New York