Practical stereoselective synthesis of an α-trifluoromethyl-α- alkyl epoxide via a diastereoselective trifluoromethylation reaction

Article abstract of DOI:10.1021/jo061839l

A practical stereoselective synthesis is reported for an α-trifluoromethyl-α-alkyl epoxide (1), which is an important pharmaceutical intermediate. The key step involves a chiral auxiliary-controlled asymmetric trifluoromethylation reaction for the introdu

Full text of DOI:10.1021/jo061839l

SCHEME 1
Practical Stereoselective Synthesis of an
r-Trifluoromethyl-r-alkyl Epoxide via a
Diastereoselective Trifluoromethylation Reaction
Jinhua J. Song,* Zhulin Tan, Jinghua Xu,
Jonathan T. Reeves, Nathan K. Yee, Ranjit Ramdas,
Fabrice Gallou, Katie Kuzmich, Lisa DeLattre, Heewon Lee,
Xuwu Feng, and Chris H. Senanayake
Department of Chemical DeVelopment, Boehringer Ingelheim
Pharmaceuticals, Inc., 900 Old Ridgebury Road/P. O. Box 368,
Ridgefield, Connecticut 06877-0368
SCHEME 2
jsong@rdg.boehringer-ingelheim.com
ReceiVed September 6, 2006
tertiary alcohol was introduced through a chiral sulfoxide-
mediated asymmetric addition to a trifluoromethyl ketone
intermediate (Scheme 1).2a However, the diastereoselectivity of
the addition step was only 2.1:1. Separation of two diastereomers
required chromatography, which is not suitable for multikilo-
gram-scale synthesis.
To further support our drug discovery programs, we needed
to develop a more efficient route to chiral epoxide 1. During
this study, we have discovered a diastereoselective, auxiliary-
controlled CF3 addition reaction to chiral keto esters. On the
basis of this discovery, a scalable, cost-effective asymmetric
route to compound 1 was achieved. In this paper, we describe
this work in a full account.3
In our new retrosynthetic analysis (Scheme 2), the target
molecule (1) would be derived from the chiral diol intermediate
2, which in turn could be synthesized by reduction of chiral
R-trifluoromethyl-R-hydroxy acid derivatives (3, R ) H).
Very few methods exist in the literature for the asymmetric
synthesis of chiral tertiary R-CF3-substituted R-hydroxy acids.
Recently, asymmetric Friedel-Crafts4 and ene reactions5 be-
tween electron-rich arenes/alkenes and trifluoropyruvates have
been used to synthesize a number of chiral tertiary R-CF3-
substituted R-hydroxy acid derivatives. Certain chiral tertiary
R-CF3-substituted R-hydroxy acids were obtained from racemic
starting materials via enzymatic reactions or chemical resolu-
tions.6
A practical stereoselective synthesis is reported for an
R-trifluoromethyl-R-alkyl epoxide (1), which is an important
pharmaceutical intermediate. The key step involves a chiral
auxiliary-controlled asymmetric trifluoromethylation reaction
for the introduction of the unique trifluoromethyl-substituted
tertiary alcohol stereogenic center in the target molecule. The
fluoride-initiated CF3 addition to chiral keto ester 6a
proceeded with a diastereoselectivity up to 86:14. The major
diastereomer was readily obtained with a >99.5:0.5 dr
through a simple crystallization of the crude product mixture.
Trifluoromethyl-containing molecules have found wide ap-
plications in the discovery and development of novel therapeutic
agents.1 For example, R-trifluoromethyl-R-alkyl epoxides were
recently utilized as key intermediates for the synthesis of
biologically active compounds such as nonsteroidal gestagens
and antiinflammatory agents.2 The greatest challenge we had
to overcome for the synthesis of this family of molecules was
the construction of the trifluoromethyl-substituted tertiary
alcohol stereogenic center. These compounds were initially
prepared in racemic forms, and the enantiomers were separated
by preparative chiral HPLC or chemical resolution. Recently,
an asymmetric route to compound 1 and its analogues was
established in which the configuration of the CF3-substituted
An unexplored approach to this class of compounds is the
direct addition of a CF3 group to keto esters in an asymmetric
(4) (a) Lyle, M. P. A.; Draper, N. D.; Wilson, P. D. Org. Lett. 2005, 7,
901. (b) Toeroek, B.; Abid, M.; London, G.; Esquibel, J.; Toeroek, M.;
Mhadgut, S. C.; Yan, P.; Prakash, G. K. S. Angew. Chem., Int. Ed. 2005,
44, 3086.
(1) Olah, G. A., Chambers, R. D., Prakash, G. K. S., Eds. Synthetic
Fluorine Chemistry; John Wiley & Sons: New York, 1992.
(2) (a) Lee, T. W.; Proudfoot, J. R.; Thomson, D. S. Bioorg. Med. Chem.
Lett. 2006, 16, 654. (b) Proudfoot, J. R.; Regan, J. R.; Thomson, D. S.;
Kuzmich, D.; Lee, T. W.; Hammach, A.; Ralph, M. S.; Zindell, R.; Bekkali,
Y. PCT Int. Appl., WO 2004063163 A1 20040729 CAN 141:140466 AN
2004:606445, 2004. (c) Lehmann, M.; Schollkopf, K.; Strehlke, P.; Heinrich,
N.; Fritzemeier, K.; Muhn, H.; Krattenmacher, R. (Schering Aktiengesell-
schaft, Germany), PCT Int. Appl., WO 98-EP3242 19980602 CAN 130:
52321 AN 1998:794993, 1998.
(3) Our preliminary account of this study has been documented in a U.S.
patent application filed in 2003 (PCT publication in 2005): Song, J. J.;
Tan, Z.; Yee, N. K.; Senanayake, C. H.; Xu, J.; Gallou, F. (Boehringer
Ingelheim Pharmaceuticals, Inc.), US 2005209488, A1 20050922.
(5) (a) Mikami, K.; Aikawa, K.; Kainuma, S.; Kawakami, Y.; Saito, T.;
Sayo, N.; Kumobayashi, H. Tetrahedron: Asymmetry 2004, 15, 3885. (b)
Aikawa, K.; Kainuma, S.; Hatano, M.; Mikami, K. Tetrahedron Lett. 2004,
45, 183.
(6) Recent examples: (a) Kuko, Y.; Yamada, K.; Nishizaki, K.; Hidaka,
T. (Mitsubishi Gas Chemical Co., Ltd., Japan), Jpn. Kokai Tokkyo Koho,
CODEN: JKXXAF JP 2005023055 A2 20050127 Application: JP 2003-
27098.3 20030704, 2005. Priority: CAN 142:155646 AN 2005:72783. (b)
Shaw, N. M.; Naughton, A.; Robins, K.; Tinschert, A.; Schmid, E.; Hischier,
M.-L.; Venetz, V.; Werlen, J.; Zimmermann, T.; Brieden, W.; de Ried-
matten, P.; Roduit, J.-P.; Zimmermann, B.; Neumueller, R. Org. Process
Res. DeV. 2002, 6, 497.
10.1021/jo061839l CCC: $37.00 © 2007 American Chemical Society
Published on Web 12/09/2006
292
J. Org. Chem. 2007, 72, 292-294

SCHEME 3
SCHEME 4
SCHEME 5
fashion. It has been known for many years that CF3 addition to
carbonyl groups can be efficiently initiated by fluorides such
as TBAF.7 However, the development of a generally applicable
asymmetric version of trifluoromethyl addition reactions has
met with very limited success.8 Chiral cinchonine-based am-
monium fluoride was first reported by the Iseki group, and it
gave moderate enantioselectivities (15-51% ee).9 Later, Caron
et al. further optimized the structure of cinchonine-derived
ammonium fluoride and achieved up to a 92% enantiomeric
excess for one single aryl ketone substrate.10 However, much
lower enantioselectivities were observed for other generic
substrates. Therefore, despite the aforementioned efforts, the
stereoselective addition of a trifluoromethide anion to carbonyl
groups still presents a major synthetic challenge.
Our strategy is to investigate the direct CF3 addition reactions
to chiral keto esters as a new method to prepare chiral tertiary
R-CF3-substituted R-hydroxy acid derivatives as shown in
Scheme 3. It is well-known that the addition of an organome-
tallic reagent to R-keto esters bearing a chiral auxiliary, such
as in the conversion of keto ester 7 to chiral R-hydroxy ester 8,
usually proceeds with high levels of stereoselectivities thanks
to the tight chelation control (Scheme 3).11 However, it was
not obvious at the outset of our study what stereochemical
outcome we would observe in the proposed asymmetric tri-
fluoromethide addition to chiral R-keto esters because no
chelation is possible for the fluoride-mediated trifluoromethy-
lation reactions. In this paper, we would like to report that this
proposed approach was indeed effective, leading to the discovery
of a diastereoselective trifluoromethylation reaction. Our pre-
liminary account on this work has previously been documented
in a patent application filed in 2003.3 During the course of the
preparation of this manuscript, two interesting papers dealing
with this topic appeared in the literature in early 2006. Pedrosa
and co-workers reported an asymmetric CF3 addition reaction
to glyoxalate-based substrates.12 Mukaiyama group reported a
synthesis of Mosher’s acid derivatives using a chiral auxiliary-
controlled asymmetric trifluoromethylation reaction of aryl-
substituted R-keto esters.13 These reports prompted us to disclose
our own findings in this area immediately.
three steps (Scheme 4). Sulfuric acid catalyzed Friedel-Crafts
alkylation of 4-fluoroanisole with methallyl chloride gave
compound 9 in 75% yield after purification by vacuum
distillation. The reaction was quite clean with 95% conversion
and less than 5% dialkylated byproduct. Chloride 9 was then
converted into its corresponding Grignard reagent. The Grignard
reagent was then allowed to react with diethyl oxalate to give
the monoester, which was hydrolyzed immediately to afford
keto acid 5 (80% overall yield from 9). At this point, all
impurities that were generated in preceding steps could be easily
removed through an acid-base extraction. Compound 5 is
typically isolated with greater than 99% purity by HPLC.
Two commonly used chiral auxiliaries, 8-phenyl menthol and
trans-2-phenylcyclohexanol, were initially selected to test the
proposed CF3 addition reactions. It was found that esterification
of keto acid 5 with 8-phenyl menthol was unsuccessful under
a variety of reaction conditions, presumably due to steric
reasons.14 On the other hand, the (1R,2S)-trans-2-phenylcyclo-
hexanol can be readily attached to keto acid 5 in refluxing
toluene in the presence of a catalytic amount of TsOH,
furnishing ester 6a in excellent yield (Scheme 5).
The trans-2-phenylcyclohexanol is commercially available
in small quantities. For larger amounts, it can be readily prepared
through enantioselective opening of cyclohexene oxide by
phenyllithium under the catalysis of a chiral Schiff base,
following a literature procedure.15 This compound was synthe-
sized in our pilot plant in multikilogram quantities using this
one-step protocol. The trans-2-phenylcyclohexanol is a highly
crystalline compound and has a relatively small molecular
weight. These properties indeed make this molecule a much
more desirable chiral auxiliary for large-scale synthesis than
8-phenyl menthol which is an oil, is very expensive, and requires
three steps to make.
Results from the investigation of the key asymmetric CF3
addition reaction are summarized in Table 1. When keto ester
6a was subjected to a standard TBAF-initiated trifluoromethy-
lation reaction in THF, a diastereoselectivity of 73:27 was
achieved with compound 10 as the major isomer (entry 1).16
When the reaction was carried out in toluene with 5 mol %
TBAF or TBAT (Bu4NPh3SiF2) at -20 °C to room temperature,
a diastereoselectivity of up to 86:14 was obtained for our
substrate (entries 6 and 7). Other solvents that were tested, for
example, DMF, methylene chloride, MTBE, and hexane, gave
To test the proposed asymmetric CF3 addition reaction, we
synthesized the requisite keto acid 5 from 4-fluoroanisole in
(7) Prakash, G. K. S.; Yudin, A. K. Chem. ReV. 1997, 97, 757.
(8) For recent reviews, see: (a) Ma, J.; Cahard, D. Chem. ReV. 2004,
104, 6119. (b) Shimizu, M.; Hiyama, T. Angew. Chem., Int. Ed. 2005, 44,
214. (c) Mikami, K.; Itoh, Y.; Yamanaka, M. Chem. ReV. 2004, 104, 1.
(9) Iseki, K.; Nagai, T.; Kobayashi, Y. Tetrahedron Lett. 1994, 35, 3137.
(10) Caron, S.; Do, N. M.; Arpin, P.; Larivee, A. Synthesis 2003, 11,
1693.
(11) (a) Whitesell, J. K.; Allen, D. E. J. Am. Chem. Soc. 1988, 110,
3585. (b) Whitesell, J. K. Chem. ReV. 1992, 92, 953.
(12) Pedrosa, R.; Sayalero, S.; Vicente, M.; Maestro, A. J. Org. Chem.
2006, 71, 2177.
(13) Kawano, Y.; Kaneko, N.; Mukaiyama, T. Chem. Lett. 2006, 35,
304.
(14) 8-Phenyl menthol ester formation through the acid chloride of
compound 5 was attempted, and a low conversion was observed (20-25%).
(15) Oguni, N.; Miyagi, Y.; Itoh, K. Tetrahedron Lett. 1998, 39, 9023.
(16) The absolute stereochemistry of the major diastereomer (10) was
confirmed by converting 10 to epoxide 1 and comparison with an authentic
sample by chiral HPLC and optical rotation.
J. Org. Chem, Vol. 72, No. 1, 2007 293

TABLE 1. Asymmetric Trifluoromethylation of Keto Ester 6a
SCHEME 6
entry
solvent
initiator (loading)
time (min)
dr
1
2
3
4
5
6
7
8
THF
DMF
TBAF (5 mol %)
TBAF (5 mol %)
TBAF (5 mol %)
TBAF (5 mol %)
TBAF (5 mol %)
TBAF (5 mol %)
TBAT (5 mol %)
TBAF (0.5 mol %)
20
20
20
20
20
20
30
20
73:27
47:53
58:42
79:21
75:25
84:16
86:14
77:23
CH2Cl2
MTBE
hexane
toluene
toluene
toluene
auxiliary and provides up to a 6.3:1 (86:14) diastereoselectivity.
The major diastereomer was readily isolated with 99.5:0.5 dr
via simple crystallization. The chiral auxiliary was quantitatively
recovered and reused. Overall, this new synthesis is practical,
chromatography free, robust, and cost effective. It has been
successfully implemented on kilogram scales.
lower selectivities (entries 2-5). Attempts were made to reduce
the amount of TBAF to minimize the amount of desilylated
product. However, this resulted in a slight drop in diastereose-
lectivity (entry 8). Therefore, 5 mol % of TBAF was used for
scale-up. It is interesting to note that reaction temperature does
not have a major influence on the stereoselectivity. The reaction
can be run at as high as -20 °C. In fact, this is a very attractive
feature for industrial applications because, unlike in regular
laboratories, it is usually difficult to achieve temperatures below
-20 °C in a production plant.
The two diastereoisomers could be separated by silica gel
column chromatography. However, for large-scale synthesis, a
direct crystallization method must be developed. A number of
solvents (THF, toluene, hexane, dichloromethane, MeOH) were
screened in the crystallization study, and it turned out that the
desired isomer 10 could be easily crystallized from the MeOH
solution of the crude product mixture. This procedure was
successfully executed to produce 3 kg of compound 10 (>99.5:
0.5 dr, >99% purity by HPLC, 50% isolated yield).
Optically pure ester 10 can be reduced directly to diol 2.
However, the separation of diol 2 and the cleaved chiral auxiliary
necessitates silica gel chromatography. To avoid this issue, ester
10 was first saponified to give hydroxy acid 3. The chiral
auxiliary was separated and recovered in nearly quantitative
yield through simple acid-base extractions (Scheme 6). The
recovered trans-2-phenylcyclohexanol can be reused without
any additional purification. For reduction of hydroxy acid 3 to
diol 2, a number of different reagents (BH3-THF, BH3-DMS,
Red-Al, NaBH4-H2SO4, NaBH4-I2, DIBAL) were evaluated. It
was found that treatment of hydroxy acid 3 with DIBAL at room
temperature led to a complete and clean conversion to diol 2
within 3 h. Other methods gave either incomplete reaction or
large amounts of byproducts. Ring closure of diol 2 was
achieved by means of MsCl/TEA to furnish chiral epoxide 1 in
99% yield with excellent purity (99.5:0.5 er, 98% by HPLC).
A batch of 2.5 kg of chiral epoxide 1 was synthesized using
this new sequence, and it proved to be identical to an authentic
sample by HPLC, 1H NMR, 13C NMR, chiral HPLC, and optical
rotation.
Experimental Section
Diastereoselective Trifluoromethylation Reaction. (RR)-5-
Fluoro-2-methoxy-γ,γ-dimethyl-R-(trifluoromethyl)-R-[(trimeth-
ylsilyl)oxy]-benzenebutanoic Acid [(1R,2S)-2-Phenylcyclohexyl]
Ester (10). Keto ester 6a (10 g, 24.2 mmol) was placed in a dry
250 mL three-neck flask equipped with a mechanical stirrer and
under a constant nitrogen flush. Anhydrous toluene (100 mL) was
added to form a slightly yellow clear solution. TMSCF3 (5.2 g,
36.4 mmol) was added, and the solution was cooled to -20 °C.
TBAF (1.2 mL, 1.2 mmol of a 1 M solution in THF) was added
slowly while the internal temperature was kept below -15 °C. The
resulting solution was stirred at -20 °C for 5 min before it was
allowed to warm to room temperature. After 30 min, HPLC
indicated that there was no ester left and that the desired CF3 adducts
were formed in an 84:16 diastereomeric ratio. Water (20 mL) was
added to the flask, and the mixture was then stirred for 15 min.
The layers were separated, and the organic layer was washed with
20 mL of brine. Solvent was removed under vacuum, and the oil
was chased with 40 mL of heptane to yield 14.2 g of crude product
as an orange oil. Under vigorous agitation, 30 mL of MeOH was
added. After about 5 min, white crystals appeared. After 15 min,
the solids were filtered and washed with MeOH (5 mL twice) and
dried. The desired isomer 10 (6.75 g, 50%) was obtained as a white
crystalline compound. The diastereomeric ratio was >99.5:0.5, and
1
the chemical purity was >99% (HPLC peak area at 220 nm). H
NMR (CDCl3, 400 MHz) δ 7.34-7.20 (m, 5H), 6.77 (m, 1H), 6.66
(m, 1H), 4.92 (ddd, J ) 4.4, 10.8, 10.8 Hz, 1H), 3.73 (s, 3H), 3.28
(d, J ) 14.6 Hz, 1H), 2.75 (ddd, J ) 3.7, 11.0, 12.4 Hz, 1H), 1.62
(dq, J ) 3.6, 12.8 Hz, 1H), 1.55-1.27 (m, 4H), 0.85 (s, 3H), 0.43
(s, 3H), -0.22 (s, 9H). 13C NMR (CDCl3, 100 MHz) δ 168.7, 166.7,
154.4, 152.6, 141.8, 137.8, 137.7, 127.3, 126.3, 125.5, 124.1, 113.2,
113.0, 110.7, 110.6, 110.5, 110.2, 79.0, 78.8, 78.5, 54.1, 48.0, 38.8,
35.0, 32.2, 29.8, 28.6, 24.2, 23.5, 23.1, 0.0. IR (cm-1) 3377, 2941,
1748, 1494, 1243, 1268, 1173, 1081, 1031, 1007, 868, 845, 741.
HRMS: Sodium adduct was observed. [M + Na]+: obsd, 577.2371;
calcd, 577.2367. Mp: 89.0 °C. [R]25 +4.34 (c 1.2, MeOH).
365
In summary, we have developed an eight-step asymmetric
synthesis of the chiral R-trifluoromethyl-R-alkyl epoxide 1,
which is an important intermediate for the synthesis of
pharmaceutically active compounds. The new synthesis features
a newly discovered chiral auxiliary-controlled asymmetric
trifluoromethylation reaction. The key asymmetric step uses the
readily available trans-2-phenylcyclohexanol as the chiral
Supporting Information Available: Experimental procedures
and characterization data for compounds 1-3, 5, 6a, and 9 as well
1
as copies of H NMR spectra for compounds 1-3, 5, 6a, 9, and
10. This material is available free of charge via the Internet at
http://pubs.acs.org.
JO061839L
294 J. Org. Chem., Vol. 72, No. 1, 2007