A short and efficient synthesis of (R,R)-2-methylcyclopropanecarboxylic acid

Article abstract of DOI:10.1021/op060249m

We report herein a short and efficient synthesis of (R,R)-2- methylcyclopropanecarboxylic acid via a Horner-Wadsworth-Emmons reaction involving commercially available (S)-propylene oxide and triethylphosphonoacetate (TEPA). The TEPA/base/propylene oxide stoichiometry was found critical to achieve high yields. We therefore studied the TEPA anion formation and stability using in situ IR spectroscopy. The reaction yield is strongly influenced by the counterion and solvent, whereas high diastereoselectivities are always obtained. Under the best experimental conditions (HexLi/MeTHF/150 °C), crude (R,R)-2-methylcyclopropanecarboxylic acid is obtained in 85-90% yield with >98% trans selectivity.

Full text of DOI:10.1021/op060249m

Organic Process Research & Development 2007, 11, 689−692
A Short and Efficient Synthesis of (R,R)-2-Methylcyclopropanecarboxylic Acid
Laurent Delhaye,† Alain Merschaert,*,†,§ Pieter Delbeke,‡ and Willy Brioˆne‡
Chemical Product Research and DeVelopment and Analytical Sciences Research and DeVelopment, Eli Lilly & Company,
Lilly DeVelopment Centre S.A., Rue Granbonpre 11, B-1348 Mont-Saint-Guibert, Belgium
Abstract:
We report herein a short and efficient synthesis of (R,R)-2-
methylcyclopropanecarboxylic acid via a Horner-Wadsworth-
Emmons reaction involving commercially available (S)-propy-
lene oxide and triethylphosphonoacetate (TEPA). The TEPA/
base/propylene oxide stoichiometry was found critical to achieve
high yields. We therefore studied the TEPA anion formation
and stability using in situ IR spectroscopy. The reaction yield
is strongly influenced by the counterion and solvent, whereas
high diastereoselectivities are always obtained. Under the best
experimental conditions (HexLi/MeTHF/150 °C), crude (R,R)-
2-methylcyclopropanecarboxylic acid is obtained in 85-90%
yield with >98% trans selectivity.
Figure 1.
Figure 2. Major impurities formed in the reaction of (S)-
propylene oxide with TEPA anions.
Introduction
It has been known for a long time that cyclopropane
derivatives can be prepared from epoxides via the Horner-
Wadsworth-Emmons (HWE) reaction.1 In recent years, this
method has successfully been extended to the preparation
of enantiomerically pure trans-2-aryl-cyclopropanecarboxylic
acid derivatives.2
Several methods have been reported to prepare chiral
nonracemic 2-methylcyclopropanecarboxylic acid.3 However,
to the best of our knowledge, all these methods have major
disadvantages with respect to the development of a large
scale process: multistep synthesis,3a,b,d,e moderate enantio-
meric excess,3a-c,g use of expensive chiral auxiliaries,3d,e,g and
consistently low overall yields.
Results and Discussion
In contrast to the reaction of styrene oxide or related aryl
epoxides with anions of triethylphosphonoacetate (TEPA),
the utilization of propylene oxide is much more challenging
from several points of view: (i) alkyl epoxides typically
afford only moderate yields of the corresponding cyclopro-
panes;4 (ii) the low boiling point of propylene oxide may
cause technical difficulties unless appropriate pressurized
equipment is used;5 (iii) a mixture of cis- and trans-
diastereomers is likely to be obtained due to the potentially
weaker steric repulsion between the methyl group and the
carbethoxy function.
From the first screening that was perfomed at a 1 mmol
scale we restricted the choice of bases to potassium, sodium
or lithium alkoxides, alkyl lithiums, and sodium hydride for
the subsequent work. Indeed, very low yields were observed
with weaker mineral (Cs2CO3, K2CO3, K3PO4, KOH) or
organic (LDA, DBU) bases. We also decided to focus on
ethereal solvents as low conversions were obtained in
hydrocarbons (toluene, xylenes).6
Herein, we wish to report the development of a short and
high-yielding synthesis of (R,R)-2-methylcyclopropanecar-
boxylic acid 2b from commercially available (S)-propylene
oxide 1 (Figure 1).
* To whom correspondence should be addressed. E-mail: alain.merschaert@
ucb-group.com.
† Chemical Product Research and Development.
‡ Analytical Sciences Research and Development.
§ Present affiliation and address: UCB Pharma S.A., Chemical Process
Development and Industrialisation, Chemin du Foriest, B-1420 Braine-l’Alleud,
Belgium.
During our initial studies, we also noticed the formation
of several major byproducts (Figure 2). Impurities 3 and 4
are derived from either a competitive protonation of the
alkoxide 6 or carbanion 8 intermediates illustrated in the
mechanism proposed below (Figure 3).2b,3f Impurity 5 results
(1) (a) Wadsworth, W. S., Jr.; Emmons, W. D. J. Am. Chem. Soc. 1961, 83,
1733. (b) Denney, D. B.; Vill, J. J.; Boskin, M. J. J. Am. Chem. Soc. 1962,
84, 3944.
(2) (a) Armstrong, A.; Scutt, J. N. Org. Lett. 2003, 5, 2331. (b) Armstrong, A.;
Scutt, J. N. Chem. Commun. 2004, 510. (c) Singh, A. K.; Rao, M. N.;
Simpson, J. H.; Li, W.-S.; Thornton, J. E.; Kuehner, D. E.; Kacsur, D. J.
Org. Process Res. DeV. 2002, 6, 618.
(3) (a) Sakaguchi, K.; Mano, H.; Ohfune, Y. Tetrahedron Lett. 1998, 39, 4311.
(b) Arai, I.; Mori, A.; Yamamoto, H. J. Am. Chem. Soc. 1985, 107, 8254.
(c) Wang, M.-X.; Feng, G.-Q. J. Mol. Catal. B: Enzym. 2002, 18, 267. (d)
Vallgarda, J.; Hacksell, U. Tetrahedron Lett. 1991, 32, 5625. (e) Vallgarda,
J.; Appelberg, U.; Csoeregh, I.; Hacksell, U. J. Chem. Soc., Perkin Trans.
1 1994, 461. (f) Walser, P.; Renold, P.; N’Goka, V.; Hosseinzadeh, F.;
Tamm, C. HelV. Chim. Acta 1991, 74, 1941. (g) Ebers, R.; Kellogg, R. M.
Recl. TraV. Chim. Pays-Bas 1990, 109, 552.
(4) (a) See ref 1. (b) Deno, N. C.; Billups, W. E.; LaVietes, D.; Scholl, P. C.;
Schneider, S. J. Am. Chem. Soc. 1970, 92, 3700.
(5) Under our optimal conditions (HexLi/MeTHF/150 °C), a pressure of 10
bar is obtained with a reactor filled at one-third of its volume. The use of
t-BuONa at 110 °C allows lowering the pressure at 4 to 5 bar.
(6) Low conversions in hydrocarbons were attributed to the formation of very
thick slurries resulting in bad mixing.
10.1021/op060249m CCC: $37.00 © 2007 American Chemical Society
Published on Web 06/13/2007
Vol. 11, No. 4, 2007 / Organic Process Research & Development
•
689

Figure 3. Proposed mechanism for the reaction of (S)-propylene oxide with TEPA anions.
Table 1. Effects of Key Reaction Parameters on Yield (2a)
h) have shown that propylene oxide is quickly converted
while only small amounts of 2a are detected together with
much larger peaks of 3, 4, and 5. This is in agreement with
a rate-limiting cyclopropane ring closure and the probable
formation of a stable chelate intermediate (Figure 4). This
latter could explain the high diastereoselectivities obtained
despite relatively moderate steric effects.
In this regard, we investigated the possibility of complet-
ing the TEPA anion addition at 75 °C (Na) or 100 °C (Li)
prior to heating the reaction mixture to 110 °C (Na) or
150 °C (Li). However, this modification did not result in
any improvement of the yield.
Although only moderate yields were obtained with
alkoxides, a dramatic improvement was observed with alkyl
lithiums. An up to 95% yield was indeed obtained with hexyl
lithium at 150 °C. We therefore assume that the presence of
tert-butanol in the experiments involving tert-butoxides leads
to the formation of increased levels of impurities 3 and 4 by
protonation of intermediates 6 and 8 as observed initially
with protic bases. It is also worth mentioning that thick
slurries (consisting likely of sodium- or lithium-diethylphos-
phate) were observed at the end of all the reactions. However,
no negative impact on the outcome of the reactions was
evidenced as indicated by the high conversions obtained.
Noteworthy, lower yields were obtained with NaOtBu upon
increasing the temperature (entries 1, 6-7). In the experi-
ments performed at 130 and 150 °C, we noticed the
formation of several unidentified byproducts which could
explain the drop of yield.8 Fortunately, most of these
unknown compounds were not observed in the experiments
involving lithium-TEPA, although the reactions were run at
150 °C to achieve a complete conversion in 16-18 h.
Obviously, this elevation in temperature also had an impact
on the reaction pressure. Pressures up to 10 bar were
measured in MeTHF using a reactor filled at one-third of its
maximum capacity. In this regard, the use of dioxane or
diglyme as higher boiling point solvents would have been
beneficial to lower the reaction pressure, but we decided to
move forward with MeTHF as it gives the highest yields
and allows a much easier recovery of 2b after hydrolysis
and Diastereomeric Ratio (2b)a
2a
2b trans:cis
(yield %)
entry
base
solvent T (°C) yield (%)
1
2
3
4
5
6
7
8
NaOtBu MeTHF
NaOtBu MTBE
NaOtBu THF
110
110
110
110
110
130
150
110
80
150
150
150
150
170
150
150
63
53
36
53
47
51
47
48
30
64
84
87
95
92
77
78
97.1:2.9 (60)
96.8:3.2
97.4:2.6
98.0:2.0
97.4-2.6
97.2:2.8
97.1:2.9
97.5:2.5 (40)
95.0:5.0
NaOtBu dioxane
NaOtBu diglyme
NaOtBu MeTHF
NaOtBu MeTHF
NaH
MeTHF
MeTHF
9
KOtBu
10
11
12
13
14
15
16
LiOtBu MeTHF
98.4:1.6
BuLi
BuLi
HexLi
HexLi
HexLi
HexLi
MeTHF
MeTHF
MeTHF
MeTHF
dioxane
diglyme
98.6:1.4 (80)
98.7:1.3 (82)
98.2:1.8 (90)
98.0:2.0 (85)
98.9:1.1
98.0:2.0
a Reaction conditions: 1.0 equiv of TEPA, 1.0 equiv of base, 1.0 equiv of
propylene oxide in 2 mL of solvent/mmol of propyelene oxide in a 160 mL Parr
stainless steel reactor with overhead stirring at the indicated temp for 18 h. A
heating rate of 10 °C/min was used except for entry 12 (3 °C/min).
from the displacement of an ethoxy group from the phos-
phorous atom of 7.
In addition, in our hands, the use of ethereal solvents
appeared critical to achieve high yields although the use of
xylenes has been reported in the literature.2a
Larger scale experiments were then performed in 160-
mL stainless steel Parr reactors equipped with overhead
stirrers. The yields of ester 2a were determined by 1H NMR
using naphthalene as an internal standard.7 A typical 1H NMR
spectrum of a crude sample is available as Supporting
Information. The diastereomeric and enantiomeric excesses
were determined by chiral GC on the isolated acid 2b.
As shown in Table 1, with the exception of the potassium
anion of TEPA (entry 9), high diastereoselectivities were
obtained in all experiments. Counterion (Li > Na . K) and,
to a lesser extent, solvent (MeTHF > diglyme ≈ dioxane ≈
MTBE > THF) have a significant effect on the yield. In-
process controls at the beginning of some reactions (1 and 2
(7) For the first experiments, the yields of 2a were cross-checked by GC and
1H NMR on isolated material. The differences typically observed (∼5%)
were attributed to the volatility of 2a.
(8) GC- and LC-MS analyses on isolated 2a indicated the possible formation
of di- and oligomers from byproducts 3, 4, and 5. Traces of tert-butyl ester
were also detected.
690
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development

Although isolation of ester 2a is easily achievable by
distillation,9 we found it much more convenient to directly
hydrolyze ester 2a to acid 2b using aqueous NaOH at
reflux for 4 h. In this regard, the use of MeTHF offers the
additional advantage of allowing the elimination of se-
veral byproducts when separating the organic layer from the
water-soluble sodium salt of 2b.10 After acidification with
concentrated HCl, the free acid 2b was extracted with
isopropylacetate. Alternatively, toluene or MTBE can be
used.
Figure 4. Chelation effect proposed to explain the high
diastereoselectivities.
Under our current best experimental conditions, isolated
yields ranging from 85 to 90% were obtained (up to 5% 2b
is lost in the aqueous layers). The quality of 2b at this stage
was sufficient for further processing in the subsequent step
of the synthesis of a compound in clinical development.11 If
needed, further purification of 2b can also be achieved by
known methods such as distillation4b or crystallization as
quinine salt.12 It is worth mentioning that quinine has
previously been used for the optical resolution of racemic
trans-2-methylcyclopropanecarboxylic acid. This latter me-
thod was initially compared with the present “chiral pool”
synthesis. However it was not judged competitive as it
affords optically pure trans-2-methylcyclopropanecarboxylic
acid in only 25% yield after four crystallizations in ace-
tone.12 Alternatively, we have found that 2b can be crystal-
lized as its dicyclohexylamine salt in isopropylacetate with
70% overall yield from (S)-propylene oxide (>99.5% ee,
>98% de).
In conclusion, we have developed a short and efficient
synthesis of (R,R)-2-methylcyclopropanecarboxylic acid from
commercially available (S)-propylene oxide and trieth-
ylphosphonoacetate. Although unoptimized, isolated yields
of 2b range from 85 to 90% under the current best
experimental conditions. Depending on the level of purity
needed in subsequent steps, 2b can either be used crude or
(9) Bp ) 50 °C/40 mbar.
(10) MeTHF solubility in water at 20 °C ) 14 g/100 g; water solubility in MeTHF
at 20 °C ) 4 g/100 g. For a study of MeTHF as efficient process solvent,
see: Aycock, D. F. Org. Process Res. DeV. 2007, 11, 156.
Figure 5. In situ IR monitoring of TEPA anion formation.
(11) Typical experimental procedure: Under a nitrogen atmosphere, n-hexyl-
lithium (2.3 M in hexanes, 8 mL, 0.0184 mol) is added dropwise over 20
min to triethylphosphonoacetate (4.5 g, 0.0197 mol) in 40 mL of anhydrous
2-methyltetrahydrofuran keeping the temp between 19 and 25 °C. After 30
min, propylene oxide (1.17 g, 0.0202 mol) is added, and the mixture is
transferred into a 160 mL Stainless Steel Parr reactor equipped with a 14
bar rupture disk (we never charged the reactor above one third of its volume
for the present procedure). The mixture is heated to 150 °C within 15 min
and stirred at this temp for 16 hours at which time 1H NMR analysis of the
crude mixture with naphthalene as internal standard indicates >95% molar
yield (see 1H NMR spectra and GC chromatograms as Supporting
Information). Water (50 mL) and 30% NaOH (25 mL) are added, and the
biphasic mixture is stirred at reflux for 5 hours. The layers are separated,
and the organic phase is discarded. Aqueous 37% HCl (25 mL) is added to
the aqueous layer, and the mixture is extracted with isopropylacetate (2 ×
50 mL). The organic layer is washed with 10% NaCl (3 × 25 mL) and
evaporated under a vacuum to yield 1.70 g of (R,R)-2-methylcyclopropan-
ecarboxylic acid (2b) as a colorless oil (ca. 95% chemical purity by 1H
NMR; >99.5% ee and >97.0% de by chiral GC; see 1H NMR spectra and
GC chromatograms as Supporting Information). Further purification of 2b
can be achieVed by crystallization as dicyclohexylamine salt: To the above
solution of crude 2b in isopropylacetate is added dicyclohexylamine (3g,
0.0165 mol). A heavy precipitation is observed after 5 min, and the
suspension is stirred overnight at 20 °C. The crystals are filtered, washed
with isopropylacetate (4 mL), and dried at 40 °C under a vacuum (3.4 g,
66% overall yield, >99.5% ee and >98% de by chiral GC; see 1H NMR
spectra and GC chromatograms as Supporting Information).
(Vide infra). On the other hand, inexpensive and nonhazard-
ous sodium tert-butoxide is also worth considering as a
potential substitute for alkyllithium.
In the preliminary studies, we had observed that the
TEPA/base/propylene oxide stoichiometry is critical to
achieve high yields. Indeed, although a large excess of TEPA
anion can be used without a significant drop of yield, an
excess of strong base causes a fast polymerization of
propylene oxide. On the other hand, with a substoichiometric
amount of base, the excess TEPA protonates the opened
intermediate preventing cyclopropane ring formation and
leading to an incomplete conversion of propylene oxide. As
such a narrow stoichiometry adjustment is sometimes chal-
lenging with strong bases. We used in situ IR spectroscopy
to monitor the anion formation (Figure 5). This methodology
allowed a better understanding of the fast kinetics of the
TEPA anion formation (“dose-controlled”) and an easy
assessment of the reaction completion and TEPA anion
stability.
(12) Bergman, R. G. J. Am. Chem. Soc. 1969, 91, 7405.
Vol. 11, No. 4, 2007 / Organic Process Research & Development
•
691

purified by various methods. The process has recently been
scaled-up to prepare 6.25 kg of 2b.13
edged for their support in the initial phase of the safety
assessment and in situ IR spectroscopy studies.
Supporting Information Available
Acknowledgment
Analytical data for the synthesis of 2b, safety studies, and
React-IR spectra. This material is available free of charge
via the Internet at http://pubs.acs.org.
We thank Mr. Jeffry K. Niemeier, CPR&D Engineering
Consultant, for his support in the safety assessment. Mr.
Geoffroy Geldhof and Dr. Alfio Borghese are also acknowl-
(13) Safety studies (adiabatic T° rise calculation and VSP) have been performed
before moving forward with the scaleup (6.25 kg). Details are included as
Supporting Information
Received for review November 28, 2006.
OP060249M
692
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development