Model study for a general approach to morphine and noroxymorphone via a rare heck cyclization

Article abstract of DOI:10.1021/ol991135g

matrix presented ss-Bromoethylbenzene was converted to the pentacycle 6 in 11 steps. The key features of this model were the toluene dioxygenase catalyzed production of diol 13 and the stereospecific Heck closure of 5 to 6.

Full text of DOI:10.1021/ol991135g

ORGANIC
LETTERS
1999
Vol. 1, No. 13
2085-2087
Model Study for a General Approach to
Morphine and Noroxymorphone via a
Rare Heck Cyclization
Dean A. Frey, Caiming Duan, and Tomas Hudlicky*
Department of Chemistry, UniVersity of Florida, GainesVille, Florida 32611-7200
hudlicky@chem.ufl.edu
Received October 9, 1999
ABSTRACT
â-Bromoethylbenzene was converted to the pentacycle 6 in 11 steps. The key features of this model were the toluene dioxygenase catalyzed
production of diol 13 and the stereospecific Heck closure of 5 to 6.
Several years ago we reported the synthesis of a complete
morphinan skeleton via a radical cyclization of aryl halide
1.1 The pentacyclic ketone 2 was obtained as shown in the
ent series, as a consequence of the configuration at C5 and
C9 (morphine numbering) in the precursor 1. The radical
cyclization of intermediates already containing the C9-C14
bond cannot be successful in producing the correct C14
stereochemistry. (Compare this approach with that of Parker3
in which the C14 is set correctly as a consequence of a
conformationally more flexible system.)
We considered the Heck reaction4 for this closure, despite
the fact that cyclizations onto highly substituted olefins are
rare. To our knowledge there is only one example of a
successful Heck cyclization onto a tetrasubstituted olefin
related to isoquinoline derivatives such as 1: that of 3. Cheng
successfully performed the Heck cyclization of 3 (X ) Br,
48%; X ) I, 72%), but the olefin was generated at C14-C9
(morphine numbering). When C9 was functionalized as a
lactam, the â-hydride elimination furnished C8-C14 unsat-
uration.5 Overman employed the Heck closure onto a
trisubstituted olefin in 4 during his recent synthesis of
cyclization produced the incorrect stereochemistry at C14
because of unfavorable conformation of the radical prior to
hydrogen abstraction, and it is clear that the same fate would
be expected for similar closures performed in the natural
series from precursors having the opposite configuration at
C5 and C9. These results, fully disclosed in a recent
publication,2 indicated that the approach based on radical
(1) Butora, G.; Fearnley, S. P.; Gum, A. G.; Stabile, M. R.; Abboud, K.
Tetrahedron Lett. 1996, 37, 8155.
(2) Butora, G.; Hudlicky, T.; Fearnley, S. P.; Stabile, M. R.; Gum, A.
G.; Gonzalez, D. Synthesis 1998, 665.
10.1021/ol991135g CCC: $18.00 © 1999 American Chemical Society
Published on Web 11/25/1999

"
"
Figure 1. Heck cyclization strategy for morphinan synthesis.
morphine.6 On the basis of this precedent we felt that the
issue of C14 stereochemistry could be successfully resolved
by using the Heck closure on intermediates such as 1.
Furthermore, â-hydride elimination of the organopalladium
intermediate would “return” the necessary oxidation state to
ring C of morphine, allowing control of C14 at that stage.
The strategy is illustrated in Figure 1. We chose isoquino-
line derivative 5, which has the “natural” configuration at
C9, with the expectation that the Heck cyclization of this
substrate could produce 6 as well as its regioisomer because
of the steric requirements for syn â-hydride elimination and
the syn arrangement of the organopalladium intermediate and
the hydrogen at C9.
In the actual applications to either 7 or 11, no regioisomers
would be expected because the C9 center will always remain
trans to the organopalladium functionality. Conversion of 1
Scheme 1
2086
Org. Lett., Vol. 1, No. 13, 1999

to 10 or 7 to 8 would eventually ensure the correct C14
stereochemistry, once the C10-C11 closure, performed by
previously established methods,1,2 is completed. The synthesis
of 5 was accomplished as shown in Scheme 1 and followed,
with few adjustments, the previously reported synthesis of
dibenzoate 14.7
and convert it to the reduced neopine intermediate 9. Rice
has demonstrated successful conversion of codeine, via
neopinone, to 14-hydroxycodeinone.12
Biooxidation of (2-bromoethyl)benzene (12)8 with Es-
cherichia coli JM109 (pDTG601) followed by the reduction
of the less substituted double bond with diimide afforded
diol 13 in 80% yield. Esterification of the two hydroxyl
groups with DCC/PhCO2H in CH2Cl2 followed by the
displacement of bromine with oxazolidine-2,4-dione in the
presence of tetramethylguanidine (TMG) in THF furnished
the protected diol 14 in 77% yield. Following reduction of
the more reactive amide carbonyl with NaBH4,9 N-acylimin-
ium ion-olefin cyclization (AlCl3 in CH2Cl2), and the
elimination of the alkyl chloride with DBU in DMSO under
reflux afforded the cyclized product 15.10 The deprotection
of the dibenzoate groups with LiOH in MeOH followed by
the selective protection of the distal (homoallylic) hydroxyl
group furnished the monoprotected diol 16. Mitsunobu
inversion of the secondary alcohol with 6-bromo-2-meth-
oxyphenol provided the Heck reaction substrate 5. Heck
cyclization of the sterically hindered tetrasubstituted olefin
proceeded in the presence of tetrakis(triphenylphospine)-
palladium(0) catalyst with Proton Sponge in refluxing toluene
for 48 h to give the tetracyclic product 6 as the only
identifiable product in a 57% yield along with 15% recovered
starting material.11
We have shown that the Heck reaction provides stereo-
and regiospecifically the neopine-type intermediate even in
the case of 5 where there is a possibility of alternate
â-hydride elimination. In both the natural and the ent series
such a possibility is removed. Therefore, olefin 8, to be
obtained from the “natural” precursor 7, should be easily
converted to neopine 9 (R ) H) and then to either 17(X )
H) or 18 (X ) OH) by an oxidation/isomerization or
epoxidation/oxidation sequence, respectively,11 and can
therefore be converted to both morphine and noroxymor-
phone. Studies on the realization of this protocol are ongoing
and will be reported in due course.
Acknowledgment. The authors thank TDC Research,
Inc., NSF (CHE-9315684 and CHE-9521489), and the U.S.
Environmental Protection Agency (R826113) for financial
support of this work.
The successful preparation of 6 allows the formulation of
a new approach to both morphine and noroxymorphone since
the issue of C10-C11 closure has been previously solved
by the acid-catalyzed cyclization.1,2 What remains to com-
plete the synthesis of both enantiomeric series is the
application of a double Mitsunobu inversion to produce 7
OL991135G
(10) In a study to determine the regiochemistry of elimination, the two
isomers of intermediate halides from the cyclization of 14, 19, and 21 were
subjected to elimination conditions with neat DBU. Isomer 19 gave
selectively 20 in 97% yield, while isomer 21 produced enol ester 22 as a
major product.
(3) Parker, K. A.; Fokas, D. J.J. Am. Chem. Soc. 1992, 114, 9688. Parker,
K. A.; Fokas, D. J. J. Org. Chem. 1994, 59, 3927.
(4) Heck, R. F. Org. React. 1982, 27, 345. Crisp, G. P. Chem. Soc. ReV.
1998, 27 (6),. 427.
(5) Cheng, C. Y.; Liou, J. P.; Lee, M. J. Tetrahedron Lett. 1997, 38,
4571.
(6) Hong, Y. H.; Kado, N.; Overman, L. E. J. Am. Chem. Soc. 1993,
115, 11028.
(7) Bottari, P.; Endoma, M. A.; Hudlicky, T.; Ghiviriga, I.; Abboud, K.
A. Collect. Czech. Chem. Commun. 1999, 64, 203.
(8) Stabile, M. R.; Hudlicky, T.; Meisels, M. L. Tetrahedron: Asymmetry
1995, 6, 537.
(9) Endoma, M. A.; Butora, G.; Claeboe, C. D.; Hudlicky, T. Tetrahedron
Lett. 1997, 38, 8833.
(11) All compounds were fully characterized by physical and spectral
data.
(12) Coop, A.; Rice, K. Tetrahedron 1999, 55, 11429.
Org. Lett., Vol. 1, No. 13, 1999
2087