A practical enantioselective synthesis of (S)-3-hydroxytetradecanoic acid

Article abstract of DOI:10.1016/S0957-4166(99)00290-6

(S)-3-Hydroxytetradecanoic acid 1 has been synthesized in an overall yield of 27% from (S)-epichlorohydrin 2 as follows: (1) regio and chemoselective epoxide opening of 2 with a Grignard reagent under the catalysis by Cu(I) followed by consecutive epoxide

Full text of DOI:10.1016/S0957-4166(99)00290-6

Pergamon
Tetrahedron: Asymmetry 10 (1999) 2945–2950
A practical enantioselective synthesis of
(S)-3-hydroxytetradecanoic acid
Keisuke Matsuyama and Masaya Ikunaka ∗
The 1st Laboratories, Research and Development Center, Nagase & Co., Ltd, 2-2-3 Murotani, Nishi-ku, Kobe 651-2241, Japan
Received 15 June 1999; accepted 8 July 1999
Abstract
(S)-3-Hydroxytetradecanoic acid 1 has been synthesized in an overall yield of 27% from (S)-epichlorohydrin 2
as follows: (1) regio and chemoselective epoxide opening of 2 with a Grignard reagent under the catalysis by Cu(I)
followed by consecutive epoxide formation; (2) regioselective epoxide opening of (S)-1,2-epoxytridecane 4 with
cyanide anion under pH controlled conditions followed by consecutive nitrile hydrolysis with alkaline H2O2 gave
crude 1; (3) its purification via the N,N-dicyclohexylammonium salt 6. The method thus devised is practical and
scalable for the industrial production of 1. © 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction
Being a natural component of lipid A, the endotoxic principle of lipopolysaccharide (LPS) embedded
in the cell surface of Gram-negative bacteria, (R)-3-hydroxytetradecanoic acid has attracted much
attention from the community of synthetic chemists as well as glycobiochemists.1,2 Being incorporated
into therapeutic agents featuring the structural motif of lipid A,2 its antipode 1 is reported to be able to
enhance the immunoadjuvant activity inherent in lipid A (Fig. 1).
Figure 1.
Therefore, we devised a practical, scalable approach to both enantiomers of 3-hydroxytetradecanoic
acid. In the present communication, it is illustrated with emphasis on the industrial viability as applied to
the synthesis of 1, the pharmaceutically more intriguing enantiomer.2
∗
Corresponding author. Tel: (81) 78 992 3163; fax: (81) 78 992 1050; e-mail: masaya.ikunaka@nagase.co.jp
0957-4166/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00290-6

2946
K. Matsuyama, M. Ikunaka / Tetrahedron: Asymmetry 10 (1999) 2945–2950
2. Results and discussion
The starting point of our synthesis was optically active epichlorohydrin (Scheme 1), both enantiomers
of which were commercially available with high enantiomeric excesses (>99.9% ee) from industrial
sources such as Chirex (US) and Daiso (Japan). It was subjected to the Cu(I)-catalyzed3 chemoselective
nucleophilic displacement by a Grignard reagent. When n-decylmagnesium bromide was prepared in a
mixed solvent system of toluene:THF (10:9), its concentration could be increased to as high as 1.8 M,4
and more beneficially, in the presence of a catalytic amount of CuI, the less hindered epoxide terminus of
(S)-epichlorohydrin 2 could be opened expeditiously at a temperature as high as 5–8°C, the range being
industrially viable and easily controllable as well. It was shown by capillary GLC (Shimadzu CBP1) that
an unidentified by-product formed incipiently eventually turned into 12-tricosanol in five hours. Being
less volatile, the latter turned out to be easily separated by distillation from the recyclized product, (S)-
1,2-epoxytridecane 4, after the base treatment as described below.
Scheme 1. Reagents and conditions: (a) 1.1 equiv. n-C10H21MgBr, 0.9 mol% CuI, toluene:THF (29:18), 3–15°C, 5 h; (b) 48%
NaOH aq., MeOH:toluene (19:39), rt, 4.5 h, 55% from 2; (c) 1.2 equiv. NaCN aq., MeOH, H2SO4, pH 10.8–12.2, 47–52°C,
1.8 h; (d) (i) H2O2 aq., NaOH aq., MeOH, reflux, 7.5 h; (ii) HCl aq., 74% from 4; (e) (c-C6H11)2NH, MeCN:MeOH (6:1), 0°C,
68%; (f) (i) 0.36N H2SO4 aq., rt, 0.5 h; (ii) 0.50×10−4N H2SO4 aq., rt, 0.5 h, overall 98%
After quenching the Grignard reaction, the resulting toluene solution was diluted with MeOH and
treated with an aqueous solution of NaOH to convert (S)-1-chloro-2-tridecanol 3 completely into (S)-1,2-
epoxytridecane 4. Essentially pure 4 was then obtained by distillation in vacuo in an overall yield of 55%
from 2.
Before moving on to the next step, the enantiomeric excess of 4 was determined as follows. Its
enantiomeric excess being too difficult to assess in a direct manner, 4 was converted into (S)-1-methoxy-
2-tridecanol 7 with NaOMe in MeOH according to the method reported by Ohta and Tetsukawa
1
(Scheme 2).5 The 1-methoxy-2-ol 7 was then analyzed by H NMR spectroscopy in the presence of
Eu[(+)-tfc]3. No splitting was observed with the signal due to the CH3O group of 7 under conditions
where the racemate, prepared separately from racemic epichlorohydrin, showed the significant chemical
shift difference (∆∆δ). Taking into account the error of measurement inherent in 1H NMR spectroscopy,
it was concluded that the enantiomeric excess of 4 was not less than 97% ee.
Having confirmed that the stereochemical integrity did not deteriorate during the Grignard reaction,
we then attempted the C1 homologation with cyanide anion. When the epoxide 4 was treated with NaCN
in aqueous alcohol under the usual conditions,6 a complex mixture resulted, probably due to the basicity
increasing with the progress of the reaction. Thus, in the course of the reaction, the pH was kept strictly

K. Matsuyama, M. Ikunaka / Tetrahedron: Asymmetry 10 (1999) 2945–2950
2947
Scheme 2. Reagents and conditions: (a) NaOMe, MeOH, reflux, 2 h, 43%
adjusted between 10.8 and 12.2 with concd H2SO4,7 which eventually allowed β-hydroxynitrile 5 to be
secured pure enough to be employed directly in the next step.
Hydrolysis of the penultimate intermediate 5 proceeded successfully, only when effected through the
8
action of alkaline H2O2 in aqueous MeOH. Since 5 was delivered as a methanolic solution after the
epoxide opening with cyanide anion, it should be natural to assume that the hydrolysis step in question
could be telescoped into the preceeding step. Indeed, these two reactions could be conducted successfully
in a seamless way to give crude 1 in an overall yield of 74% from 4.
Before entering into purification of 1 thus obtained, its enantiomeric excess was assessed to see if 1
could survive the relatively harsh conditions applied without significant racemization. According to the
method of Keegan et al.,9 crude 1 was converted into p-bromophenacyl ester 8, and the latter analyzed
by HPLC on Chiralpak® AS (Scheme 3). By comparing the chromatograms between 8 and its racemate
prepared from racemic 4, the enantiomeric excess of 8 was determined reliably as 96.6%. This implied
that the telescoped conversion of 4 into 1 only led to a nominal racemization from >97% ee to 96.6% ee.
Scheme 3. Reagents and conditions: (a) 2,40-dibromoacetophenone, Et3N, EtOAc, rt, 19.5 h, quant.
Finally, crude 1 was purified by way of its crystalline salt with N,N-dicyclohexylamine 6 according
to the literature precedent.10 A single crystallization of 6 from MeCN:MeOH (6:1) succeeded in
significantly upgrading both chemical and enantiomeric purity. The salt 6 thus purified in a yield of
68% was treated with H2SO4 to liberate the free acid 1. As a result, (S)-3-hydroxytetradecanoic acid 1
was obtained as crystals in 99.5% ee and 98% yield from 6, and its overall yield from (S)-epichlorohydrin
2 amounted to 27%.
3. Conclusion
We have devised a scalable, industrially viable approach to (S)-3-hydroxytetradecanoic acid 1, which
dispenses with both tedious chromatographic purification and expensive asymmetric catalyst. Our
methodology has enabled an expeditious synthesis of 1 starting from commercially available (S)-
epichlorohydrin 2 to give 1 with 99.5% ee in an overall yield of 27%.
4. Experimental
4.1. General
1H NMR spectra were recorded on a Varian UNITY-400 spectrometer at 400 MHz by using CDCl3 as
a solvent and tetramethylsilane as an internal standard. FT-IR spectra were recorded on a Perkin–Elmer

2948
K. Matsuyama, M. Ikunaka / Tetrahedron: Asymmetry 10 (1999) 2945–2950
1600 spectrometer. Mass spectral analyses were performed with a Hewlett Packard 5890 series II
(GC) and an HP5971A (MS). High-resolution mass spectra were recorded on a Jeol JMS-DX-303
spectrometer. Elemental analyses were performed with a Perkin–Elmer 240 C analyzer. Optical rotations
were measured with a Horiba SEPA-200. Melting points were measured with an Electrothermal 1A8104.
Melting points (mps) and boiling points (bps) were uncorrected.
4.2. (S)-1,2-Epoxytridecane 4
Under an atmosphere of N2, 1,2-dibromoethane (0.040 ml, 0.46 mmol) was added to a suspension of
Mg turnings (2.5 g, 0.10 mol) in dry THF (18 ml). To the water-cooled mixture was added a solution of
1-bromodecane (22 g, 98 mmol) in dry toluene (20 ml) dropwise at a gentle reflux over a period of 20
min. The mixture was then stirred at ambient temperature for an additional 30 min. Under an atmosphere
of N2, the Grignard reagent solution thus prepared (ca. 55 ml) was added dropwise to an ice-cooled
mixture of CuI (0.15 g, 0.79 mmol), (S)-epichlorohydrin 2 [Chirex Inc.; >99.9% ee assured by chiral
GLC (Supelco γ-DEX™ 225; 30 m×0.25 mm œ; carrier gas: He, 93 ml/min; oven temperature: 70°C;
detection: FID); 8.3 g, 90 mmol], and dry toluene (5.0 ml) at about 3°C over a period of 3 h. After
the addition was complete, the dropping funnel used was rinsed with dry toluene (4.0 ml). The reaction
mixture was stirred at a temperature between 5°C and 8°C for 5 h, during which the progress of the
reaction was monitored by GLC [Shimadzu CBP1; 25 m×0.25 mm œ; carrier gas: He, 111 ml/min; oven
temperature: 40–250°C (10°C/min) +250°C (19 min); detection: FID]. After 12N HCl aqueous solution
(10 ml) was added, the mixture was concentrated in vacuo to remove THF.
To the residue was added water (20 ml), and layers were separated. The aqueous layer was extracted
with toluene (5.0 ml×2). To the combined organic layer and toluene extracts were added MeOH (19 ml)
and 48% NaOH aqueous solution (8.0 ml). The mixture was stirred at ambient temperature for 4.5 h, and
the completion of the reaction was confirmed by running the GLC analysis under the same conditions as
mentioned above. After evaporating MeOH in vacuo, layers were separated, and the aqueous layer was
extracted with toluene (5.0 ml×2). The combined organic layer and toluene extracts were washed with
water (5.0 ml), dried over MgSO4 (3.0 g), and concentrated in vacuo. The residue was distilled in vacuo
to give 4 (9.7 g) in 55% yield from 2 as a colorless oil: bp 104–107°C/4–5 mmHg (lit.11 123–125°C/13
mmHg); [α]21 −10.7 (c 1.17, Et2O) {lit.11 [α]21 −11.0 (c 1.20, Et2O)}; 1H NMR δ 0.88 (3H, t, J=6.8
D
D
Hz), 1.20–1.30 (16H, m), 1.40–1.50 (2H, m), 1.50–1.60 (2H, m), 2.46 (1H, dd, J1=5.1 Hz, J2=2.8 Hz),
2.75 (1H, dd, J1=5.1 Hz, J2=3.9 Hz), 2.91 (1H, ddt, J1=5.5 Hz, J2=3.9 Hz, J3=2.8 Hz); IR ν (film) 3042
(w), 2925 (s), 2854 (s), 1466 (m), 1410 (w), 1378 (w), 1260 (w), 1129 (w), 917 (w), 836 (m), 722 (w)
cm−1; MS (70 eV) m/z 198 (M+).
4.3. Determination of the enantiomeric excess of 4
To a solution of 4 (0.08 g, 0.40 mmol) in dry MeOH (1.0 ml) was added 28% NaOMe in MeOH (0.1
ml), and the homogeneous mixture was stirred and heated at reflux for 2 h. 1.0N HCl aqueous solution
(1.0 ml) was added, and then MeOH was evaporated in vacuo. The residue was extracted with Et2O (1.0
ml×2). The combined Et2O extracts were dried over MgSO4 (0.1 g) and concentrated in vacuo to give 7
(0.04 g, 43%): 1H NMR δ 0.88 (3H, t, J=7.0 Hz), 1.20–1.40 (18H, m), 1.40–1.50 (2H, m), 3.23 (1H, dd,
J1=9.6 Hz, J2=8.0 Hz), 3.39 (3H, s), 3.42 (1H, dd, J1=9.6 Hz, J2=6.4 Hz), 3.70–3.80 (1H, m).
(ꢀ)-1-Methoxy-2-tridecanol rac-7 was prepared from (ꢀ)-epichlorohydrin via (ꢀ)-1,2-epoxytridecane
in the same manner as described above.
When 1H NMR spectroscopy was observed with rac-7 in the presence of 0.3 molar equiv. of tris[(+)-
3-(trifluoromethylhydroxymethylene)camphorato]europium(III), the singlet due to its OCH3 group split

K. Matsuyama, M. Ikunaka / Tetrahedron: Asymmetry 10 (1999) 2945–2950
2949
into two singlets: 4.66 ppm for 4; and 4.53 ppm for its antipode. However, no such splitting was detected
with 7 under the same 1H NMR conditions as above. Thus, taking into account the intrinsic error of 1H
NMR spectroscopy, the enantiomeric excess of 7 was determined to be not less than 97% ee.
4.4. N,N-Dicyclohexylammonium (S)-3-hydroxytetradecanoate 6
To a solution of 4 (>97% ee; 1.9 g, 9.7 mmol) in MeOH (8.0 ml) was added 29% (w/w) aqueous
solution of NaCN (1.97 g, 11.6 mmol). The reaction mixture was stirred at a temperature between 48°C
and 52°C for 1.8 h during which time its pH was maintained strictly between 10.8 and 12.4 by adding
97% H2SO4 (0.40 g, 4.0 mmol). The completion of the reaction was confirmed by GLC under the same
conditions as applied in the preparation of 4.
White inorganic precipitates were filtered off, and washed with MeOH (18 ml). To the combined filtrate
and washing were added 30% aqueous solution of H2O2 (2.0 ml) and 48% aqueous solution of NaOH
(2.0 ml). The mixture was stirred and heated at reflux for 10 h, during which time the progress of the
reaction was monitored by TLC [Merck Kieselgel 60; n-hexane:AcOEt:AcOH (20:10:1)]. The mixture
was allowed to cool to ambient temperature, and 2.0N HCl aqueous solution (20 ml) was added. After the
mixture was ice-cooled with stirring, solids precipitated were collected by filtration, washed with water
(5.0 ml), and dried at an oven temperature of 40°C to give crude 1 (1.8 g) as a pale yellow solid.
This was dissolved in MeOH (1.8 ml). To the solution was added a solution of N,N-dicyclohexylamine
(1.3 g, 7.1 mmol) in MeCN (10 ml). After the mixture was ice-cooled with stirring, solids precipitated
were collected by filtration, washed with ice-chilled MeCN (6.0 ml), dried at an oven temperature of
40°C to give 6 (2.1 g, 51% from 4) as white crystals: mp 97.5–98.5°C [lit.12: mp 94–95°C for the (R)
isomer]; [α]20 +4.97 (c 1.99, MeOH) {lit.12: [α]20 −4.4 (c 2.0, MeOH) for the (R) isomer}.
D
D
4.5. Determination of the enantiomeric excess of 6
1.0N HCl aqueous solution (2.0 ml) was added to 6 (0.05 g, 0.10 mmol) at an ambient temperature. The
suspension was stirred, and extracted with Et2O (5.0 ml). The Et2O extract was concentrated in vacuo
to give free acid 1 as white crystals. This was dissolved in AcOEt (3.0 ml). To the solution were added
2,40-dibromoacetophenone (0.05 g, 0.2 mmol) and Et3N (0.03 ml, 0.2 mmol). The resulting solution was
then stirred at ambient temperature for 19.5 h. The inorganic precipitates were filtered off, and washed
with hot AcOEt (2.0 ml). The combined filtrate and washing were concentrated in vacuo to give 8 (0.07
1
g, quant.) as white crystals: H NMR δ 0.88 (3H, t, J=6.8 Hz), 1.20–1.40 (20H, br. s), 2.57 (1H, dd,
J1=15.4 Hz, J2=9.2 Hz), 2.69 (1H, dd, J1=15.4 Hz, J2=3.2 Hz), 3.30 (1H, br. s), 4.10–4.20 (1H, m), 5.32
(1H, d, J=16.4 Hz), 5.43 (1H, d, J=16.4 Hz), 7.65 (2H, d, J=8.4 Hz), 7.79 (2H, d, J=8.4 Hz).
Racemic 8 was also prepared in the same manner as described above, and its structural identity was
confirmed by 1H NMR.
HPLC on Chiralpak® AS [elution: n-hexane:2-propanol (97:3) at a flow rate of 1.0 ml/min; tempera-
ture: 35°C; detection: UV at 254 nm)] gave a complete separation between the enantiomers of rac-8: tR
23.0 min for 8; 26.2 min for its antipode. HPLC analysis of 8 under the same conditions as described
above showed its enantiomeric excess to be 99.4% ee.
4.6. (S)-3-Hydroxytetradecanoic acid 1
After 6 (4.4 g, 99.5% ee) was suspended in 0.36N H2SO4 (35 ml), the mixture was stirred at ambient
temperature for 30 min. Solids precipitated were collected by filtration and washed with water (30 ml)
to give crude 1 as white crystals. This was suspended again in 5.0×10−4N H2SO4 (20 ml) and stirred

2950
K. Matsuyama, M. Ikunaka / Tetrahedron: Asymmetry 10 (1999) 2945–2950
at ambient temperature for 30 min. The solids were collected by filtration, washed with water (15 ml),
and dried at an oven temperature of 40°C to give 1 (2.5 g) in 98% yield from 6: mp 72.8–73.1°C (lit.10:
25
mp 71.8°C); [α]25 +16.0 (c 1.03, CHCl3) {lit.10: [α] +16.1 (c 1, CHCl3)}; H NMR δ 0.88 (3H, t,
1
D
D
J=7.2 Hz), 1.20–1.40 (18H, m), 1.40–1.60 (2H, m), 2.47 (1H, dd, J1=16.5 Hz, J2=9.0 Hz), 2.58 (1H,
dd, J1=16.5 Hz, J2=3.2 Hz), 4.00–4.10 (1H, m); IR ν (KBr) 3554 (s), 2919 (s), 2846 (s), 1676 (s), 1469
(s), 1433 (m), 1394 (m), 1361 (w), 1294 (s), 1258 (m), 1225 (s), 1068 (m), 939 (m), 874 (m), 720 (w),
524 (m) cm−1; HRMS (200 eV; CI) 245.2135, calcd for C14H28O3 [(M+H)+] 245.2118. Anal. calcd for
C14H28O3: C, 68.81; H, 11.55; found: C, 68.80; H, 11.61. The enantiomeric excess of 1 thus obtained was
determined to be 99.5% ee after analyzing 8 [2.66 g; prepared from 1 (1.44 g, 5.87 mmol) quantitatively]
by chiral HPLC under the same conditions as described above.
Acknowledgements
We thank Chirex Inc. for a generous gift of (S)-epichlorohydrin. We also thank Ms. Akiko Kuramura
and Ms. Yuriko Tawara of the 2nd Laboratories, Research and Development Center, Nagase & Co.,
Ltd for their skillful assistance in performing HPLC analysis. Mr. Masafumi Moriwaki, Director of
Research and Development Center, Nagase & Co., Ltd is gratefully acknowledged for his encouragement
throughout this work, and Professor Takeshi Kitahara of the University of Tokyo (Tokyo, Japan) for his
encouragement to publish it.
References
1. (a) Huang, G.; Hollingsworth, R. I. Tetrahedron: Asymmetry 1998, 9, 4113 and references cited therein. (b) Imaki, K.;
Hashimoto, S.; Wakatsuka, H. JP Patent 94,041,175, 1994; Ono Pharmaceutical Co., Ltd. (c) Shiozaki, M.; Kurakata, S.;
Tatsuta, T.; Maeda, H.; Nishijima, M. Tetrahedron 1997, 53, 16041.
2. (a) Kusumoto, S.; Fukase, K.; Suda, Y.; Oikawa, M. J. Synth. Org. Chem., Jpn. 1996, 54; Special Issue in English, 976 and
references cited therein. (b) Fukase, K.; Liu, W. C.; Suda, Y.; Oikawa, M.; Wada, A.; Mori, S.; Ulmer; A. J.; Rietschel, E.
T.; Kusumoto, S. Tetrahedron Lett. 1995, 36, 7455. (c) Liu, W. C.; Oikawa, M.; Fukase, K.; Suda, Y.; Winarno, H.; Mori,
S.; Hashimoto, M.; Kusumoto, S. Bull. Chem. Soc. Jpn. 1997, 70, 1441.
3. (a) Huynh, C.; Derguini-Boumechal, F.; Linstrumelle, G. Tetrahedron Lett. 1979, 20, 1503. (b) Rotstein, D. M.; Walker,
K. A. M. Tetrahedron: Asymmetry 1993, 4, 1521. (c) Kobayashi, Y.; Nakano, M.; Okui, H. Tetrahedron Lett. 1997, 51,
8883.
4. n-Decylmagnesium bromide is commercially available from Aldrich as a 1.0 M ether solution.
5. Ohta, H.; Tetsukawa, H. J. Chem. Soc., Chem. Commun. 1978, 849.
6. Senda, S.; Mori, K. Agric. Biol. Chem. 1987, 51, 1379.
7. Yoshimura, H.; Furukawa Y.; Takehira, Y. JP Patent 93,310,671, 1993; Daiso Co., Ltd.
8. Nicolaou, K. C.; Petasis, N. A.; Uenishi, J.; Zipkin, R. E. J. Am. Chem. Soc. 1982, 104, 5557.
9. Keegan, D. S.; Hagen, S. R.; Johnson, D. A. Tetrahedron: Asymmetry 1996, 7, 3559.
10. Nakahata, M.; Imaida, M.; Ozaki, H.; Harada, T.; Tai, A. Bull. Chem. Soc. Jpn. 1982, 55, 2186.
11. Mori, K.; Otsuka, T. Tetrahedron 1985, 41, 547.
12. Tai, A.; Nakahata, M.; Harada, T.; Izumi, Y. Chem. Lett. 1980, 1125.