Enantioselective synthesis of the carbacephem antibiotic loracarbef via Mitsunobu and Dieckmann cyclization from an unnatural amino acid

Article abstract of DOI:10.1016/S0040-4039(03)01483-7

The nucleus of the carbacephem antibiotic loracarbef was synthesized in a highly efficient and enantioselective fashion from 2S,3S-2-amino-3-hydroxy-6-heptenoic acid (AHHA), which was derived from enzyme-catalyzed condensation of glycine and 4-pentenaldehyde. The bicyclic framework of this compound was established through sequential Mitsunobu reaction and aldol condensations.

Full text of DOI:10.1016/S0040-4039(03)01483-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 5991–5993
Enantioselective synthesis of the carbacephem antibiotic
loracarbef via Mitsunobu and Dieckmann cyclization from an
unnatural amino acid
Jerry W. Misner, Jack W. Fisher, John P. Gardner, Steve W. Pedersen, Kristina L. Trinkle,
Billy G. Jackson and Tony Y. Zhang*
Chemical Process Research and Development, Lilly Research Laboratories, A Division of Eli Lilly and Company, Indianapolis,
IN 46285-4813, USA
Received 14 May 2003; revised 12 June 2003; accepted 15 June 2003
Abstract—The nucleus of the carbacephem antibiotic loracarbef was synthesized in a highly efficient and enantioselective fashion
from 2S,3S-2-amino-3-hydroxy-6-heptenoic acid (AHHA), which was derived from enzyme-catalyzed condensation of glycine and
4-pentenaldehyde. The bicyclic framework of this compound was established through sequential Mitsunobu reaction and aldol
condensations.
© 2003 Published by Elsevier Ltd.
b-Lactams occupy a unique place in the history of
heterocyclic chemistry for their role in combating infec-
tious diseases around the globe. They are still the best
selling class of pharmaceuticals by volume in the world
after more than seven decades of widespread clinical
applications. In the search for a wider antibacterial
spectrum and longer systemic effect, the carbon analog
of penicillins and cephalosporins, namely carbapenems
and carcephems, have emerged as a new generation of
b-lactams featuring enhanced chemical and serum sta-
bility over their sulfur parents. This culminated in the
launch of the first carbapenem and carbacephem anti-
biotics, imipenem (1) and loracarbef (2),1 respectively.
challenging task. With the increasing threat of newly
discovered infectious pathogens and drug resistant
strains of known ones, b-lactams remain a privileged
platform for medicinal chemists in search of inhibitors
of bacterial cell wall synthesis, and numerous disease
related proteases. Reported herein is a new methodol-
ogy for the synthesis of carbacephems via the Mis-
tunobu cyclization.
2S,3S-2-Amino-3-hydroxy-6-heptenoic acid (3, AHHA)
has been demonstrated to be available in large scale
from serine hydroxymethyltransferase (SHMT) cata-
lyzed condensation between 4-pentenaldehyde and
glycine.2 This highly functionalized material in essen-
tially enantiopure form appeared as a versatile interme-
diate for building the loracarbef (2) framework.3 Thus
the carboxyl and the a-amino groups were protected as
p-nitrobenzyl ester (pNB-) and phenoxyacetamide (V-),
affording compound 4. Attempts to obtain the b-lactam
by cyclizing this compound, however, were met with
Unlike carbapenem, no naturally occurring carba-
cephem has been reported, and hence their entire struc-
tural activity relationship (SAR) studies were based on
compounds of total or semisynthetic origin. The sensi-
tive nature of these densely functionalized chiral com-
pounds make their enantioselective synthesis
a
* Corresponding author. E-mail: zhang@lilly.com
0040-4039/$ - see front matter © 2003 Published by Elsevier Ltd.
doi:10.1016/S0040-4039(03)01483-7

5992
J. W. Misner et al. / Tetrahedron Letters 44 (2003) 5991–5993
predominant formation of oxazoline 5, either with the
alcohol under Mitsunobu conditions or the correspond-
ing mesylate in the presence of various bases.
nes for their lower cost and easier byproduct separa-
tion. Early screening of phosphites quickly ruled out
(MeO)3P due to mixed phosphite (10) formation. Tri-
ethyl phosphite was as effective as PPh3 albeit a larger
excess was required. Both (iPrO)3P and (PhO)3P
resulted in slow rates and poor conversions. The sol-
vent factor was also investigated. Acetonitrile and
methylene chloride generally required more reagents to
complete the reaction, concomitant with a large extent
of a-epimer (8) formation, as compared to THF and
toluene. Reaction temperature played a critical role
governing the selectivity. Lower temperature not only
led to slower rate, but also a greater proportion of
compound 8. Poor solubility of 6 in aromatic solvent at
lower temperature also points toward a higher reaction
temperature. It is important to note that it is essential
to have all 6 dissolved in the reaction media or risk
higher impurity content, and it is preferable to add
DIAD to a mixture of all other ingredients due to the
instability of (EtO)3P /DIAD adduct. After an extensive
screening of various reaction conditions, we found a
combination of triethyl phosphite, diisopropylazodicar-
boxylate (DIAD) and toluene as a solvent at higher
temperature afforded the best overall yield and selectiv-
ity of the desired b-lactam 7. As an interesting side
note, the threo (2S,3R) isomer of 6 gave exclusively
elimination product 9 under virtually all of these condi-
tions, providing an unexpected but useful pathway for
impurity control.
Replacing the phenoxyacetyl with benzyloxycarbonyl
offered little relief in this regard. The insufficient nucle-
ophilicity of the amide nitrogen and the proximity of
the a-amido oxygen made the five membered ring for-
mation a favorable process. This problem could be
solved by enhancing the nucleophilicy of the nitrogen
via conversion into an ionizable sulfonamate, followed
by nucleophilic displacement of the mesylate group to
form the b-lactam ring.3 However, this approach neces-
sitates multiple steps of mesylation, amide activation,
cyclization, and alkylation of the b-lactam nitrogen for
the cephem ring fusion. We chose to focus on the
Mitsunobu4 cyclization for a more direct synthesis of 2.
Formation of b-lactams via the Mitsunobu reaction has
been demonstrated first by Bose and Miller on various
systems.5 To avoid complication associated with the
deleterious a-amido nitrogen participation, the amino
group of 3 was protected as a phthalimido group to
afford 6 in 87% yield. The corresponding mesylate of 6
proved susceptible to elimination, yielding compound 9
(50%) even in the absence of a base standing at room
temperature. Initial tests under the standard Mitsunobu
conditions, however, afforded b-lactam 7 in good yield
(Scheme 1). Several issues had to be addressed in the
context of scaling up this reaction. First, the hallmark
problem associated with the traditional Mitsunobu
reaction, (i.e. removal of phosphine oxide byproduct
from the reaction stream), proved difficult too without
resorting to chromatography. Secondly, the reaction
was found to be not stereospecific, generating various
amounts of epimeric lactam 8. Towards these ends
phosphites were evaluated as a substitute for phosphi-
With compound 7 in hand, elaboration of the terminal
alkene took the facility of ozonlysis to afford carboyxlic
acid 11 in 95% yield with in situ oxidation (Scheme 2).
Conversion to the phenyl ester was accomplished using
phenyl chloroformate as both a phenol donor and
carboxylate activator, setting the stage for Dieckmann
cyclization. The cyclization is best performed using 3.3
Scheme 1.

J. W. Misner et al. / Tetrahedron Letters 44 (2003) 5991–5993
5993
Scheme 2.
References
equiv. of t-BuOLi between −8 and −15°C, affording
enol 13 in 83% yield. Alternatively, the two steps from
11 to 13 can be carried out in a one-pot fashion in
slightly higher yield (89%), furnishing the carbacephem
skeleton. Chlorination of the enol and deprotection of
the ester went uneventfully to afford compound 15.
1. For examples of selective syntheses of 2, see: (a) Evans, D.
A.; Sjogren, E. B. Tetrahedron Lett. 1985, 26, 3787; (b)
Bodurow, C. C.; Boyer, B. D.; Brennan, J.; Bunnell, C. A.;
Burks, J. E.; Carr, M. A.; Doecke, C. W.; Eckrich, T. M.;
Fisher, J. W.; Gardner, J. P.; Graves, B. J.; Hines, P.;
Hoying, R. C.; Jackson, B. G.; Kinnick, M. D.; Kochert,
C. D.; Lewis, J. S.; Luke, W. D.; Moore, L. L.; Morin, J.
M., Jr.; Nist, R. L.; Prather, D. E.; Sparks, D. L.; Vladu-
chick, W. C. Tetrahedron Lett. 1989, 30, 232.
2. Dotzlaf, J. E.; Gazak, R. J.; Kreuzman, A. J.; Kroeff, E.
P.; Queener, S. W.; Vicenzi, J. T.; Yeh, W. K.; Joseph
Martin; Zock, J. M. Eur. Pat. 628638, 1995.
3. Jackson, B. G.; Pedersen, S. W.; Fisher, J. W.; Misner, J.
W.; Gardner, J. P.; Staszak, M. A.; Doecke, C.; Rizzo, J.;
Aikins, J.; Farkas, E.; Trinkle, K. L.; Vicenzi, J.; Rein-
hard, M.; Kroeff, E. P.; Higginbotham, C. A.; Gazak, R.
J.; Zhang, T. Y. Tetrahedron 2000, 56, 5667.
Removal of the phthalimido protecting required consid-
erable optimization due to the sensitive nature of the
molecule. Deprotection could be achieved either by
enzyme catalyzed hydrolysis or assisted cleavage using
methylhydrazine to give the loracarbef nucleus 16.6
This established an efficient formal total synthesis of
the novel antibiotic 2.
In summary, we have demonstrated the utilization of
the Mitsunobu reaction and Dieckmann condensation
for the enantioselective synthesis of loracabef, starting
from an unnatural aminoacid. All of the synthetic steps
are high yielding and no chromatographic separation,
or expensive chiral auxiliaries were required.
4. For a review on Mitsunobu reaction, see: Hughes, D. L.
Org. React. 1992, 42, 336.
5. For the synthesis of b-lactams via Mitsunobu reaction, see:
(a) Bose, A. K.; Manhas, M. S.; Sahu, D. P.; Hedge, V. R.
Can. J. Chem. 1984, 62, 2498; (b) Bose, A. K.; Sahu, D. P.;
Manhas, M. S. J. Org. Chem. 1981, 46, 1229; (c) Miller,
M. J.; Morrison, M. A. J. Org. Chem. 1983, 48, 4421; (d)
Galeotti, N.; Montagne, C.; Poncet, J.; Jouin, P. Tetra-
hedron Lett. 1992, 33, 2807.
Acknowledgements
We would like to thanks Eli Lilly’s Physical Chemistry
group for assistance in characterization, and Professor
Marvin Miller for insightful discussions.
6. Jackson, B. G.; Gardner, J. P.; Heath, P. C. Tetrahedron
Lett. 1990, 31, 6317–6320.