Scaleable syntheses of isomeric limonene aziridines from the commercially available mixture of cis- and trans-limonene oxides

Article abstract of DOI:10.1021/op0498464

A short and efficient route to both isomers of limonene aziridine is described. The process is amenable to scale-up and allows easy access to multigram quantities of these highly useful chiral building blocks.

Full text of DOI:10.1021/op0498464

Organic Process Research & Development 2005, 9, 221−224
Communications to the Editor
Scaleable Syntheses of Isomeric Limonene Aziridines from the Commercially
Available Mixture of cis- and trans-Limonene Oxides
Michael V. Voronkov,* Alexander V. Gontcharov, Ramanaiah C. Kanamarlapudi, Paul F. Richardson, and Zhi-Min Wang
Lexicon Pharmaceuticals, 350 Carter Road, Princeton, New Jersey 08540, U.S.A.
Abstract:
interest for some time, the area of asymmetric aziridination
has only recently become a major focus. An alternative to
4
A short and efficient route to both isomers of limonene aziridine
is described. The process is amenable to scale-up and allows
easy access to multigram quantities of these highly useful chiral
building blocks.
this approach would be to utilize molecules from the chiral
pool and develop chemistry to transform them into optically
active aziridines. The most obvious starting materials to carry
out such transformations would be amino acids, and indeed
numerous methods have been developed for the transforma-
tion of amino acids into the corresponding optically pure
Introduction
4,5
aziridines via the corresponding amino-alcohols. Although,
this would offer simple, rapid access to a number of
diamines, we believed that the flexible nature of the amino
acids side chain would not be suitable for developing chiral
Terpenes, and in particular limonene, are an important
class of naturally occurring chiral compounds widely used
in organic synthesis either as starting materials in the
synthesis of optically pure molecules or as the chiral core
of the numerous chiral auxiliaries or asymmetric ligands
employed in enantioselective transformations. In the major-
ity of cases, the ligands derived from terpenes are amino-
alcohols as opposed to diamines. The use of diamines as
1
6
ligands. Thus, we turned our attention to the naturally
occurring terpenes, the substituents of which are in geo-
metrically defined environments as a result of their cyclic
nature. In conjunction with our programme to develop ligands
for such asymmetric transformations, as well as our general
interest in the chemistry of these naturally occurring chiral
2
3
ligands in asymmetric synthesis is well established, although
the use of the naturally occurring terpenes to provide such
ligands has been underutilized. The simplest way to access
diamines in enantiomerically pure form is from stereoselec-
tive ring opening of the corresponding chiral aziridines.
While methods for the synthesis of chiral epoxides, and thus
chiral amino-alcohols, have been an area of active research
7
building blocks, we required a scaleable, economical method
to obtain the aziridines derived from limonene in diastereo-
merically pure form. However, the stereoselective function-
alization of the endocyclic double bond of limonene has
presented a formidable challenge thus severely limiting its
application in synthesis. For example, most of the olefin
epoxidation methods, whilst displaying moderate to excellent
regioselectivity, give an almost equimolar mixture of dia-
4
*
Corresponding author. E-mail: mvoronkov@lexpharma.com. Telephone:
6
09-466-5500. Fax: 609-466-3562.
(1) For examples of syntheses utilizing limonene as the chiral template, see:
8
stereomeric limonene epoxides.
(
(
a) Van Tamelen, E. E.; Anderson, R. J. J. Am. Chem. Soc. 1972, 94, 8225.
b) Pawson, B. A.; Cheung, H.-C.; Gurbaxani, S.; Saucy, G. J. Am. Chem.
Limonene oxides have been used as key intermediates
en route to the corresponding vicinal amino-alcohols that
have also found application as ligands in asymmetric
Soc. 1970, 92, 336. (c) Baudouy, R.; Prince, P. Tetrahedron 1989, 45, 2067.
d) Paquette, L. A.; Kang, H.-J. J. Am. Chem. Soc. 1991, 113, 2610. (e)
(
Mori, K.; Kato, M. Tetrahedron Lett. 1986, 27, 981. (f) Marron, B. E.;
Nicolaou, K. C. Synthesis 1989, 537. (g) Tius, M. A.; Kerr, M. A. Synth.
Commun. 1988, 18, 1905. (h) Dauphin, G. Synthesis 1979, 799.
(5) For examples of ring closure of amino-alcohols and their derivatives, see:
(a) Kelly, J. W.; Eskew, N. L.; Evans, S. A. J. Org. Chem. 1986, 51, 95.
(b) Kuyl-Yeheskiely, E.; Lodder, G. A.; van der Marel, G. A.; van Boom,
J. H. Tetrahedron Lett. 1987, 28, 1211. (c) Wipf, P.; Miller, C. P.
Tetrahedron Lett. 1992, 33, 6267. (d) Pfister, J. R. Synthesis 1994, 969.
(6) An exception to this would be to utilize proline as the amino acid, and
derivatives of this important synthon have been widely used as chiral
auxilaries. For example, see: Soai, K.; Ookawa, A.; Tatsuya, K.; Ogawa,
K. J. Am. Chem. Soc. 1987, 109, 7111.
(
2) For examples, see: (a) Goralski, C. T.; Chrisman, W.; Hasha, D. L.;
Nicholson, L. W.; Rudolf, P. R.; Zakett., D.; Singaram. B. Tetrahedron:
Asymmetry 1997, 8, 3863. (b) Masui, M.; Shiori, T. Tetrahedron 1995, 51,
8363. (c) Masui, M.; Shiori, T. Synlett 1995, 49. (d) Kauffman, G. S.; Harris,
G. D.; Dorow, R. L.; Stone, B. R. P.; Parsons, R. L., Jr.; Pesti, J. A.; Magnus,
N. A.; Fortunak, J. M.; Confalone, P. N.; Nugent, W. A. Org. Lett. 2000,
2
, 3119. (e) Noyori, R.; Kitamura, M.; Suga, S.; Kawai, K. J. Am. Chem.
Soc. 1986, 108, 6071.
(
(
3) Togni, A.; Venanzi, L. M. Angew. Chem., Int. Ed. Engl. 1994, 33, 497.
4) For recent reviews on aziridines including their synthesis in optically active
form and their ring opening reactions, see: (a) Sweeney, J. B. Chem. Soc.
ReV. 2002, 31, 247. (b) McCoull, W.; Davis, F. A. Synthesis 2000, 1347.
(7) For a review on ClickChem, see: Kolb, H. C.; Finn, M. G.; Sharpless, K.
B. Angew. Chem., Int. Ed. 2001, 40, 2004.
(8) For examples of the epoxidation of the endocyclic double bond of limonene,
see: (a) Newhall, W. F. J. Org. Chem. 1959, 24, 1673. (b) Knoll, W.; Tamm,
C. HelV. Chim. Acta 1975, 58, 1162. (c) Suemune, H.; Kawahara, T.; Sakai,
K. Chem. Pharm. Bull. 1986, 34, 550. (d) Lange, G. L.; Neidert, E. E.;
Orrom, W. J.; Wallace, D. J. Can. J. Chem. 1978, 56, 1628. (e) White, J.
D.; Ruppert, J. F.; Avery, M. A.; Torri, S.; Nokami, J. J. Am. Chem. Soc.
1981, 103, 1813. (f) Yamasaki, M. J. Chem. Soc., Chem. Commun. 1972,
606.
(
c) Atkinson, R. S. Tetrahedron 1999, 55, 1519. (d) Jacobsen, E. N. In
ComprehensiVe Asymmetric Catalysis; Jacobsen, E. N., Pfaltz, A., Yama-
moto, H., Eds.; Springer-Verlag: Berlin, Heidelberg, New York, 1999; Vol.
2
1
, p 607. (e) Osborn, H. M. I.; Sweeney, J. B. Tetrahedron: Asymmetry
997, 8, 1693. (f) Tanner, D. Angew. Chem., Int. Ed. Engl. 1994, 33, 599.
(g) Tanner, D. Pure Appl. Chem. 1993, 1319.
1
0.1021/op0498464 CCC: $30.25 © 2005 American Chemical Society
Vol. 9, No. 2, 2005 / Organic Process Research & Development
•
221
Published on Web 02/11/2005

Scheme 1. Separation of limonene oxides
catalysis.9,10 The separation of the mixture of limonene oxide
the amino-alcohol, by quaternisation of the amine to generate
a leaving group followed by nucleophilic ring closure
(Scheme 1).
11
12
isomers by distillation is difficult, though partial separation
by distillation has been achieved on a commercial scale.
However, to separate the epoxides, there are chemical
methods of isomer resolution based on the differences in rates
of the cis-limonene oxide and trans-limonene oxide opening
Given the versatility of the pure epoxides, we envisioned
that in a similar manner the closely related terpene aziridines
also have the potential to become useful intermediates in
organic synthesis, as well as source of both chiral diamine
or aminophosphine chiral ligands. Unfortunately, this area
of terpene chemistry remains relatively unexplored. While
several methods for the preparation of terpene aziridines have
been reported, they all vary in their degree of regio- and
diastereoselectivity, and as such there is no general, efficient,
either by secondary amines9 or via hydrolysis.
,12
13,14
trans-
Diaxial ring opening of the trans-limonene oxide is more
facile, so this isomer reacts much faster than the other one,
and this difference can be exploited to separate the isomers.
After nucleophilic ring opening, simple acid-base extraction
allows the unreacted cis isomer to be separated from the
reaction. The trans-limonene oxide can then be obtained from
1
5
safe, and selective route to such systems.
Herein, we report a method for the highly selective
synthesis of either isomer of limonene aziridine starting from
the cheap, commercially available 1:1 mixture of limonene
oxides.
(
9) (a) Steiner, D.; Sethofer, S. G.; Goralski, C. T.; Singaram, B. Tetrahedron:
Asymmetry 2002, 13, 1477. (b) Chrisman, W.; Camara, J. N.; Marcellini,
K.; Singaram, B.; Goralski, C. T.; Hasha, D. L.; Rudolf, P. R.; Nicholson,
L. W.; Borofychuk, K. K. Tetrahedron Lett. 2001, 42, 5805.
(
10) For examples of other chiral auxiliaries derived from limonene, see: (a)
Comins, D. L.; Guerra-Weltzien, L.; Salvador, J. M. Synlett 1994, 972. (b)
Comins, D. L.; Guerra-Weltzien, L. Tetrahedron Lett. 1996, 37, 3807. (c)
Ho, T.-L.; Lee, K.-Y. Tetrahedron Lett. 1995, 36, 947.
11) Fractional distillation is possible but extremely inefficient. Royals, E. E.;
Leffingwell, J. C. J. Org. Chem. 1966, 31, 1937.
12) Under harsh conditions, both epoxides can be opened with dimethylamine,
and the corresponding amino-alcohols can be separated by salt formation
and crystallization. For examples of this and the susbsequent conversion of
the amino-alcohols back to the diastereomerically pure epoxides, see: (a)
Baker, R.; Borges, M.; Cooke, N. G.; Herbert, R. H. J. Chem. Soc., Chem.
Commun. 1987, 414. (b) Newhall, W. F. J. Org. Chem. 1964, 29, 185. (c)
Kuczynski, H.; Piatkowski, K. Roczniki Chem. 1959, 33, 299. (d) Patrick,
R.; Newhall J. J. Agric. Food Chem. 1960, 8, 397.
Results and Discussion
The epoxides were converted to the aziridines through
the intermediate azido-alcohols, which were cyclized by
reaction with triphenylphosphine. The transformation begins
with the opening of the limonene oxide mixture 1 with
sodium azide in the presence of ammonium chloride as the
catalyst (Scheme 2). The opening of each isomer is stereo-
selective, and the stereochemistry of the opening arising from
(
(
(
13) For examples, see: (a) Steiner, D.; Ivison, L.; Goralski, C. T.; Appell, R.
B.; Gojkovic, J. R.; Singaram, B. Tetrahedron: Asymmetry 2002, 13, 2359.
(15) For examples, see: (a) Davis, C. E.; Bailey, J. L.; Lockner, J. W.; Coates,
R. M. J. Org. Chem. 2003, 68, 75. (b) Bochvic, B.; Kapuscinski, J.;
Olejniczak, B. Roczniki Chem. 1971, 45, 869. (c) Subbaraj, A.; Rao, O. S.;
Lwowski, W. J. Org. Chem. 1989, 54, 3945. (c) Bergmeier, S. C.; Seth, P.
P. Tetrahedron Lett. 1999, 40, 6181. (d) Corey, E. J.; Ortiz de Montellano,
P. R.; Lin, K.; Dean, P. D. G. J. Am. Chem. Soc. 1967, 89, 2797. (e) Avruch,
L.; Oehlschlager, A. C. Synthesis 1973, 622. (f) van Ende, D.; Krief, A.
Angew. Chem., Int. Ed. Engl. 1974, 13, 279. (g) Parrish, E. J.; Nes, W. D.
Synth. Commun. 1988, 18, 221. (h) Nes, W. D.; Parrish, E. J. Lipids 1988,
23, 375. (i) Corey, E. J.; Riddiford, L. M.; Ajami, A. M.; Yamamoto, H.;
Anderson, J. E. J. Am. Chem. Soc. 1971, 93, 1815.
(b) Jones, J.; dos Santos, A. G.; de Lima Castro, F. Synth. Commun. 1996,
2
6, 2651. (c) Cole-Hamilton, D. J.; Salles, L.; Nixon, A. F.; Russell, N. C.;
Clarke, R.; Pogorzelec, P. Tetrahedron: Asymmetry 1999, 10, 1471. (d)
Van der Warf, M. J.; Jongejan, H.; Franssen, M. C. R. Tetrahedron Lett.
2
001, 42, 5521.
(
14) Enzyme mediated hydrolyses has also been reported. See: (a) Weijers, C.
A. G. M. Tetrahedron: Asymmetry 1997, 8, 639. (b) Van der Werf, M. J.;
Orru, R. V. A.; Overcamp, K. M.; Swarts, H. J.; Osprian, A.; de Bont, J.
A. M.; Faber, K. Appl. Microbiol. Biotechnol. 1999, 52, 38.
222
•
Vol. 9, No. 2, 2005 / Organic Process Research & Development

Scheme 2
the trans-diaxial approach of the nucleophile has long been
sion of the azido-alcohol 2b to the cis-aziridine 3b requires
elevated temperatures and proceeds smoothly in refluxing
dioxane over a period of 16 h. Both aziridine isomers, 3a
and 3b, were purified by vacuum distillation to afford the
desired products in greater than 98% purity and in high yield.
established.1
1,12b,16
As such, nucleophilic attack of trans-
limonene oxide occurs exclusively at the secondary C-2
carbon and leads to formation of 2a. Similarly, opening of
the cis-limonene oxide, leading to formation of 2b, occurs
at the tertiary carbon (C-1) and, therefore, is slower. This
rate difference can be exploited when bulkier amines are
employed as the nucleophile, as it is sufficiently large to
allow clean separation of the unreacted cis-limonene oxide
Conclusions
In conclusion, we have developed a practical, scaleable
method for the preparation of either isomer of limonene
aziridine from the commercially available mixture of li-
monene oxides. This methodology is extremely efficient and
avoids the separation of the limonene oxides (and subsequent
separate processing of each diastereomer) by either physical
or chemical methods. The key to the efficiency of the
separation is the exploitation of the differences in rate
between the two azido-alcohol diastereomers in the ring
closure.
(Scheme 2). However, the reaction of the smaller and
extremely nucleophilic azide anion is extremely rapid with
both epoxide isomers making kinetic resolution at this stage
impossible.1
5a,17
The subsequent step, the aziridine formation, involves a
18
pseudo-Staudinger reaction between the mixture of azides
a and 2b, with triphenylphosphine to effect both azide
reduction with concomitant ring closure to yield the aziri-
2
1
9
dine. Interestingly, the ring closure has previously been
achieved by conversion of the azido-alcohol to the corre-
Experimental Section
20
sponding mesylate, and reductive cyclisation, though using
such a protocol here may not allow the azido-alcohol
diastereomers to be distinguished. We found that the reac-
tivities of the two isomeric azido alcohols toward tri-
phenylphosphine differ significantly enough to allow a highly
efficient kinetic resolution of the mixture. The Staudinger
reaction of the secondary azide is much faster, so azido
alcohol 2a is converted completely to the corresponding
aziridine 3a at room temperature over a period of 48 h. The
resulting aziridine 3a is easily separated from the unreacted
azido-alcohol 2b by a simple acid-base extraction. Conver-
General Comments. NMR experiments were conducted
13
with a Bruker ARX300 spectrometer at 75 MHz for C and
1
3
00 MHz for H spectra. Samples were dissolved in CDCl
3
with TMS as the internal reference. Reactions were carried
out under nitrogen under anhydrous conditions. LC/MS
analysis was conducted on a HP1100 with a Polaris C18
A-5µ column.
Preparation of Azido-5-isopropenyl-2-methyl-cyclo-
hexanols 2a and 2b: To a solution of a commercial mixture
of (-)-cis-limonene oxide and (-)-trans-limonene oxide 1
(76 g, 0.50 mol, Acros, 1:1 mixture) in methanol (200 mL)
(
16) Deslongchamps, P. Stereoelectronic Effects in Organic Chemistry; Perga-
mon: New York, 1983; Chapter 5.
was added sodium azide (65 g, 1.0 mol) (Caution: highly
toxic chemical; handle and dispose according to OSHA and
your state regulations) and ammonium chloride (27 g, 0.50
mol). The resulting mixture was heated at reflux with
magnetic stirring for 24 h (the reaction was monitored by
GC). The mixture was allowed to cool to rt, and the solids
were removed by filtration. Methanol was removed in vacuo,
and the oily residue was diluted with ether (200 mL) and
filtered through a 1-in Celite plug. The ether solution was
dried over magnesium sulfate and concentrated to yield 96
g of a crude 1:1 mixture (98% yield; 94% purity by GC) of
17) Coates1 reports modest selectivity in the ring opening of the mixture of
aziridines at low temperature using diethylaluminium azide as the nucleo-
phile.
5a
(
(
18) Gololobov, Y. G.; Zhmurova, I. N.; Kasukhin, L. F. Tetrahedron 1981, 37,
4
37.
(
19) For a similar transformation, see: Sommerdijk, N. A. J. M.; Buynsters, P.
J. J. A.; Akdemir, H.; Geurts, D. G.; Nolte, R. J. M.; Zwanenburg, B. J.
Org. Chem. 1997, 62, 4955. Interestingly the authors report that the success
of the transformation is dependent on the scale on which the reaction is run
and that the transformation is not amenable to being run on large scale.
This is not the case with our procedure.
(
20) Ferrero, L.; Geribaldi, S.; Rouillard, M.; Azzaro, M. Can. J. Chem. 1975,
5
3, 3227.
Vol. 9, No. 2, 2005 / Organic Process Research & Development
•
223

azides 2a and 2b as yellow viscous oil, which was used
without further purification.
Preparation 4-Isopropenyl-1-methyl-7-aza-bicyclo-
[4.1.0]heptane 3b. The crude azido-alcohol 2b isolated from
the preparation of 3a was dissolved in reagent grade 1,4-
dioxane (ca. 200 mL), and the solution was heated with
triphenylphosphine (75 g, 0.28 mol) at reflux for 36 h (again
the reaction could be easily monitored by GC). Upon
completion, the reaction was allowed to cool to rt before
being diluted with ether (300 mL). The triphenylphosphine
oxide, which precipitated, was filtered off, and the organic
filtrate was washed with an aqueous solution of citric acid
(2 × 100 mL, 96 g in 200 mL of water). The aqueous layer
was separated and washed with DCM (2 × 150 mL), and
its pH was adjusted to pH 7 by addition of 1 N aq NaOH
solution. The product was extracted with ether (2 × 200 mL).
The organic extracts were dried over magnesium sulfate,
filtered, and concentrated in vacuo, and the residue was
purified by vacuum distillation (bp 55-60 °C/2 mm) to yield
Preparation 4-Isopropenyl-1-methyl-7-aza-bicyclo-
[4.1.0]heptane 3a. The crude mixture (96 g, 0.492 mol) of
2a and 2b was dissolved in 200 mL of THF, and triphen-
ylphosphine (75 g, 0.28 moL) was added. The reaction
mixture was stirred at rt for 2 days (the reaction was
monitored by GC) and then concentrated in vacuo. The
residue was dissolved in ether (200 mL) and filtered through
a cotton plug. The filtrate was washed twice with an aqueous
solution of citric acid (2 × 100 mL, 96 g in 200 mL of
water). The organic layer was separated, dried over magne-
sium sulfate, filtered, and concentrated in vacuo. The
resulting oil contained unreacted azido-alcohol, 2b, and was
used directly in the next reaction without further purification.
The aqueous layer was washed with DCM (2 × 150 mL),
and then the pH was adjusted to ca. 7 by adding NaOH
solution (1 N aq, ∼90 mL). The crude aziridine was extracted
with ether (2 × 200 mL), and the organic extracts were dried
over magnesium sulfate, filtered, and concentrated in vacuo.
The residue was distilled under vacuum (bp 60-65 °C/2 mm)
1
22 g (59%) of the aziridine 3b as a colorless oil. H (300
MHz; CDCl
3
): 1.11-1.19 (m, 1H), 1.25 (s, 3H), 1.41-1.58
(m, 1H), 1.67 (s, 3H), 1.74-1.85 (m, 3H), 1.91-2.13 (m,
13
3
3H), 4.68 (s, 1H), 4.71 (s, 1H). C (75 MHz; CDCl ): 21.4,
1
to yield 25 g (67%) of the aziridine 3a as a colorless oil. H
26.8, 27.8, 29.8, 30.8, 34.7, 36.9, 39.5, 109.2, 149.6.
(300 MHz; CDCl
3
): 1.27 (s, 3H), 1.33-1.57 (m, 4H), 1.61
(
3
3
s, 3H), 1.79 (td, 1H, J ) 10.5 Hz, 2.9 Hz), 1.93-2.07 (m,
H), 4.62 (s, 2H). 1 C (75 MHz; CDCl
3
Received for review August 10, 2004.
OP0498464
3
): 20.1, 24.6, 26.1,
0.3, 30.7, 34.8, 37.3, 41.2, 108.4, 149.3.
224
•
Vol. 9, No. 2, 2005 / Organic Process Research & Development