Efficient, stereoselective synthesis of trans-2,5-disubstituted morpholines

Article abstract of DOI:10.1021/ol049861t

(Equation presented) Enantio- and diastereoselective syntheses of trans-2,5-disubstituted morpholine derivatives are described. The routes are initiated by the reaction of enantiopure epoxides (2) with amino alcohols (3) and address the problem of regiose

Full text of DOI:10.1021/ol049861t

ORGANIC
LETTERS
2004
Vol. 6, No. 6
1045-1047
Efficient, Stereoselective Synthesis of
trans-2,5-Disubstituted Morpholines
Brian A. Lanman and Andrew G. Myers*
Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
myers@chemistry.harVard.edu
Received January 23, 2004
ABSTRACT
Enantio- and diastereoselective syntheses of trans-2,5-disubstituted morpholine derivatives are described. The routes are initiated by the
reaction of enantiopure epoxides (2) with amino alcohols (3) and address the problem of regioselective hydroxyl activation−ring closure of the
resulting amino diol adducts for (amino alcohol-derived) alkyl substituents of different steric demands.
In the course of research directed toward the solid-phase
synthesis of a library of saframycin analogues,1 we had need
of gram quantities of several trans-2,5-disubstituted mor-
pholine derivatives (1). These were required to function as
dual linkers,2 providing both a point for attachment to the
solid support (via the 2-substituent of the morpholine ring)
and a site for linkage to various R-amino aldehydes that we
employed as synthetic precursors (by amino nitrile formation
with the nitrogen atom of the morpholine ring). It was
desirable that the latter transformation be diastereoselective,
for which purpose the 5-substituent of the morpholine ring
played a critical role.3
Despite the fact that trans-2,5-disubstituted morpholine
derivatives are widely represented among pharmacologically
active substances (Figure 1), there are no commercial sources
(1) Myers, A. G.; Lanman, B. A. J. Am. Chem. Soc. 2002, 124, 12969-
12971.
(2) (a) Tam, J. P.; Tjoeng, F. S.; Merrifield, R. B. Tetrahedron Lett.
1979, 20, 4935-4938. (b) Tam, J. P.; Tjoeng, F. S.; Merrifield, R. B. J.
Am. Chem. Soc. 1980, 102, 6117-6127.
(3) Lanman, B. A.; Myers, A. G. Manuscript in preparation.
(4) (a) Berlin, A. Y.; Sycheva, T. P. Zh. Obshch. Khim. 1950, 20, 640-
647. (b) Dobrev, A.; Perie, J. J.; Lattes, A. Tetrahedron Lett. 1972, 13,
4013-4016. (c) Uozumi, Y.; Tanahashi, A.; Hayashi, T. J Org. Chem. 1993,
58, 6826-6832. (d) Sun, G.; Savle, P. S.; Gandour, R. D.; Bha´ırd, N. N.
a′; Ramsay, R. R.; Fronczek, F. R. J. Org. Chem. 1995, 60, 6688-6695.
(e) Take, K.; Konishi, N.; Shigenaga, S.; Kayakiri, N.; Azami, H.; Eikyu,
Y.; Nakai, K.; Ishida, J.; Morita, M. Preparation of Piperazines for Treating
or Preventing Tachykinin-Mediated Diseases. PCT Int. Appl. WO 00035915,
Jun 22, 2000. (f) Ancliff, R. A.; Cook, C. M.; Eldred, C. D.; Gore, P. M.;
Harrison, L. A.; Hayes, M. A.; Hodgson, S. T.; Judd, D. B.; Keeling, S. E.;
Lewell, X. Q.; Mills, G.; Robertson, G. M.; Swanson, S.; Walker, A. J.;
Wilkinson, M. Preparation of Morpholinylmethylureas as CCR-3 Antago-
nists. PCT Int. Appl. WO 03082861, Oct 9, 2003.
Figure 1. Examples of biologically active trans-2,5-disubstituted
morpholine derivatives.
of such compounds of which we are aware, and existing
protocols for their synthesis lead to diastereomeric mixtures
of products.4 Although retrosynthetic analysis of this general
target structure is not difficult (a component-based approach
10.1021/ol049861t CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/26/2004

Scheme 1. Synthesis of Morpholine 9
Figure 2. Identification of readily available chiral starting materials
for the enantio- and diastereoselective synthesis of trans-2,5-
disubstituted morpholines.
involving enantiopure epoxides 2 and amino alcohols 3 as
starting materials is readily derived; see Figure 2), we found
that implementation of this approach was not completely
straightforward. In light of the value of trans-2,5-disubsti-
tuted morpholines in different contexts, we outline here two
synthetic solutions for their preparation, optimized for
5-substituents with different steric requirements.
The chiral building blocks 2 and 3 in the proposed
component-based route (Figure 2) were selected in part
because they were readily available, the former by the
Jacobsen hydrolytic kinetic resolution procedure7 and the
latter from amino acids (and widely available from com-
mercial sources). The general sequence by which we
envisioned that these starting materials would be transformed
into the desired morpholines involved epoxide (2) opening
by the amino group of the amino alcohol (3), N-protection
of the resultant adduct, then selective hydroxyl activation,
ring closure, and N-deprotection. It proved necessary to
develop two variations on this scheme in order to access
trans-2,5-disubstituted morpholines with 5-substituents of
differing steric requirements (see Schemes 1 and 2). These
are discussed in sequence, beginning with a description of
the route to 2,5-disubstituted morpholine derivatives with
less hindered 5-substituents (illustrated with a 5-methyl
group; compound 9, Scheme 1).
Synthesis of the trans-2,5-disubstituted morpholine deriva-
tive 9 began with the reaction of the enantiopure epoxide
(S)-41 with a 4-fold excess of enantiopure D-alaninol (5) in
n-propanol at reflux, providing exclusively the product of
monoalkylation (6), with the amino group bonding to the
less-hindered carbon atom of the epoxide ring (99%).
Exposure of the resultant amino diol (6) to tosyl chloride
(1.1 equiv) in the presence of triethylamine (2.0 equiv) in
dichloromethane at 0 °C provided selectively the N-tosylated
diol 7 (77%). This product was then cyclized by a method
that we had previously reported, without specific discussion
in that context, that allows for simultaneous hydroxyl
activation-ring closure of suitably substituted 1,5-diols.1 This
one-step procedure involved deprotonation of 7 with excess
sodium hydride (2.5 equiv) in THF (0 f 23 °C), followed
by the addition of 1 equiv of p-toluenesulfonyl imidazole at
0 °C, and afforded the N-tosyl morpholine derivative 8 in
99% yield. The stereoisomeric product that would have arisen
from activation of the secondary hydroxyl group was not
detected. Cleavage of the N-protective group proceeded in
quantitative yield when the sulfonamide 8 was subjected to
treatment with excess sodium in ethanolic ammonia at -78
°C followed by warming to ambient temperature. Extractive
isolation then provided the desired morpholine derivative 9
in pure form (100% yield).
The use of the p-toluenesulfonamide protective group in
the sequence outlined in Scheme 1 was critical; when a
related sequence was attempted using instead an N-benzyl
protective group, cyclization, as above,1 led to a 1.5:1 mixture
of trans-2,5- and cis-2,6-disubstituted morpholine deriva-
tives.8 Similarly, the use of an N-carbamoyl protective group
proved to be problematic, in this case due to oxazolidinone
formation between the primary hydroxyl group and the
N-protective group. In contrast, the use of the N-tosyl group,
as outlined in Scheme 1, provided the desired morpholine
derivative 9 as a single diastereomer in four steps (75% yield)
and required only one chromatographic purification (to
separate excess tosyl chloride from the intermediate 7).9 This
represents a marked improvement over related syntheses that
were previously reported4 and has proven effective for the
preparation of 9 on a 20-g scale.10
Although the sequence summarized in Scheme 1 was
highly effective for the application illustrated, when we
(8) The undesired cis-2,6-disubstituted morpholine is presumed to have
arisen by a sequence involving reversible aziridinium ion formation.
Tosylate-mediated opening of the aziridinium ion at the more substituted
terminus followed by ring closure would proceed with net retention of
configuration at the methyl-bearing center while effecting its transfer from
the 5- to the 6-position of the cyclic product.
(9) The 2-nitrobenzenesulfonamide (nosyl) protective group (Fukuyama,
T.; Jow, C.-K.; Cheung, M. Tetrahedron Lett. 1995, 36, 6373-6374) was
found not to be a viable substitute for the N-tosyl protective group in the
synthesis of 9, as it was not stable toward the conditions of attempted
cyclization (cf. 7 f 8, Scheme 1; nosyl cleavage occurred, forming a mixture
of 2-nitrophenyl ethers as products).
(5) Kawashima, S.; Matsuno, T.; Yaguchi, S.; Sasahara, H.; Watanabe,
T. Preparation of Heterocyclic Compounds as Antitumor Agents. PCT Int.
Appl. WO 02088112, Nov 7, 2002.
(6) (a) Ong, J.; Kerr, D. I. B.; Bittiger, H.; Waldmeier, P. C.; Baumann,
P. A.; Cooke, N. G.; Mickel, S. J.; Froestl, W. Eur. J. Pharmacol. 1998,
362, 27-34. (b) Kuo, S.-C.; Blythin, D. J.; Kreutner, W. 2-Substituted
Morpholine and Thiomorpholine Derivatives as GABA-B Antagonists. U.S.
Patent 5,929,236, Jul 27, 1999.
(7) Tokunaga, M.; Larrow, J. F.; Kakiuchi, F.; Jacobsen, E. N. Science
1997, 277, 936-938.
(10) We acknowledge Dr. Jonathan White for conducting the synthesis
of 9 on a 20-g scale.
1046
Org. Lett., Vol. 6, No. 6, 2004

but led instead to O-tosylation-elimination (forming an
N-tosyl enamine), necessitating the development of an
alternative procedure. Although Mitsunobu-based cyclo-
dehydration protocols were promising in preliminary experi-
ments, the imperfect regioselectivity (primary vs secondary
hydroxyl activation) that we observed in these experiments
led us to develop an alternative two-step activation-closure
sequence that allowed for the preparation of a single
regioisomer. Thus, reaction of 14 with p-toluenesulfonyl
chloride (1.25 equiv), 4-(dimethylamino)pyridine (0.4 equiv),
and triethylamine (3.7 equiv) afforded a 4.4:1 mixture of
the primary tosylate (15) and the corresponding 1,5-bis-O-
tosylate. When this mixture was heated at reflux in a solution
of potassium carbonate in tert-butyl alcohol, the N-protected
morpholine derivative 16 was formed in 71% yield (from
14, two steps).11 Interestingly, the use of stronger bases or
aprotic solvents in this transformation afforded predominantly
the product of elimination of the O-tosyl group.12
Scheme 2. Synthesis of 2,5-Disubstitued Morpholines with
Larger 5-Substituents
Completion of the synthesis of the trans-2,5-disubstituted
morpholine 17 was achieved by the deprotection of 16 using
sodium naphthalenide in tetrahydrofuran.13 The trans-2,5-
disubstituted morpholine 17 was obtained as a single dia-
stereomer, in 77% yield (48% overall yield; seven steps).
We have used this method to prepare gram quantities of 17.
The protocol of Scheme 2 also proved to be directly appli-
cable to the synthesis of diastereomerically pure 5-benzyl
morpholine derivative 18 (40% yield; seven steps).
attempted to use the same sequence to prepare morpholine
derivatives with larger 5-substituents (phenyl, benzyl; see
compounds 17 and 18) selective O-tosylation occurred during
attempted N-protection, apparently as a result of increased
steric hindrance about the amino group. Thus, the sequence
outlined in Scheme 2 was developed.
In summary, we have described two efficient, highly
stereoselective methods for the preparation of gram quantities
of trans-2,5-disubstituted morpholine derivatives, starting
from readily available, enantiomerically pure starting materi-
als. These complementary protocols allow for the preparation
of trans-2,5-disubstituted morpholine derivatives bearing both
large and small 5-substituents. We believe that these
protocols may be of value in a range of applications.
As illustrated (Scheme 2), this sequence began in the same
way as that of Scheme 1, by the selective coupling of the
epoxide 4 with an amino alcohol, in this case enantiopure
(R)-(-)-2-phenylglycinol (10). To achieve N-tosylation in
this series, however, we found it necessary to transiently
protect the hydroxyl groups of the amino diol product (11)
using hexamethyldisilazane (2.04 equiv) in the presence of
trimethylsilyl chloride (0.4 equiv) in THF at 0 °C. The bis-
trimethylsilyl ether that was formed (12, 96%) was then
N-tosylated upon treatment with p-toluenesulfonyl chloride
(4.0 equiv) in pyridine containing triethylamine. The latter
additive was necessary to avoid premature cleavage of the
trimethylsilyl ether protective groups during this relatively
slow N-tosylation reaction (15 h required for complete
reaction). Excess p-toluenesulfonyl chloride was scavenged
by the addition of ethylenediamine, greatly facilitating the
isolation of the sulfonamide 13 (chromatography on silica
gel, 99%). The trimethylsilyl ether groups within 13 were
then readily cleaved upon treatment with sodium methoxide
in methanol, affording the cyclization precursor, diol 14
(97%).
Acknowledgment. Financial support from the NIH is
gratefully acknowledged. B.A.L. acknowledges an NSF
predoctoral fellowship.
Supporting Information Available: Experimental pro-
cedures and spectroscopic data for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL049861T
(11) The 1,5-bis-O-tosylate formed during the synthesis of 15 did not
react under the cyclization conditions; it was readily separated from 16 by
flash column chromatography on silica gel.
(12) Cooper, K. A.; Dhar, M. L.; Hughes, E. D.; Ingold, C. K.; MacNulty,
B. J.; Woolf, L. I. J. Chem. Soc. 1948, 2043-2049.
(13) Sodium naphthalenide/tetrahydrofuran was used in lieu of sodium/
ammonia in order to avoid competing Birch reduction of the aromatic ring.
Attempted cyclodehydration of 14 using the prior one-
step protocol1 (cf. 7 f 8, Scheme 1) was not successful,
Org. Lett., Vol. 6, No. 6, 2004
1047