A stereoselective synthesis of a 2-functionalized-methyl-1b-methylcarbapenem
key intermediate via decarboxylation
Woo-Baeg Choi, Jaemoon Lee, Joseph E. Lynch,*† R. P. Volante, Paul J. Reider and Robert A. Reamer
Process Research, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
TBDMSO
An efficient synthesis of a key intermediate 2a for the
O
O
H
H
N
synthesis of 2-functionalized methyl-1b-methylcarbapenem
antibiotics 1 has been realized via a stereoselective decarbox-
ylation reaction.
OAc
OBn
+
OBn
O
H
The emergence of multiple drug resistant bacteria and the
alarming increase in the number of infections resulting from
these organisms have clearly demonstrated an urgent need for
new antimicrobial agents. Methicillin resistant Staphylococcus
aureus (MRSA) is one of the more common of these infectious
agents and in the early 1990s accounted for about 20% of the
bacterial cultures isolated in hospitals.1 MRSA remains a
challenging target of antimicrobial research programs through-
out the pharmaceutical industry. 2-Functionalized-methyl-
1b-methylcarbapenem antibiotics 1 have drawn much attention
due their potent activity against a variety of pathogens including
MRSA.2 However, the lack of an efficient synthesis of these
carbapenems has hindered their development as clinical
candidates. The 2-hydroxymethyl-1b-methyl ketone 2a is a key
intermediate in the synthesis of 1 and several syntheses of 2a in
variously protected forms have appeared in the literature.2a,3
5
6
i
O
O
TBDMSO
v
TBDMSO
CH3
CH3
H
H
N
H
H
N
OBn
OBn
iii
2a
OBn
O
O
iv
R
O
R
9 R = TBDMS
2c R = H
7 R = H
8 R = TBDMS
ii
Scheme 1 Reagents and conditions: i, K2CO3, DMF, 45 °C, 1 h; ii, Et3N,
TBDMSOTf, DMF; iii, HCO2H, EtOAc, 5% Pd/C, 30 psi H2, room temp.,
1 h; iv, Bun4NF, CH2Cl2, 0 °C; v, 10% Pd/C, EtOH, 45 psi H2
observed a remarkable diastereospecificity in the decarboxyla-
tion reaction.4d We found that only the (1R)-ester 3d underwent
decarboxylation to give 2d. The (1S)-diastereomer underwent
decarboxylation only under forcing conditions resulting in ring
cleavage and giving 4b as a major product. We were concerned
that such diastereospecificity would be problematic with the 2:1
mixture of diastereomers 8. We thus separated the two isomers
of 8 (silica gel chromatography) and investigated the decarbox-
ylation of each isomer individually. Each benzyl ester 8 was
subjected to hydrogenolysis conditions (3 equiv. HCO2H, 5%
Pd/C, 30 psi, room temp., 1 h) in EtOAc. Surprisingly both
isomers cleanly gave 9 as a single stereoisomer ( > 99:1 b:a).5
Encouraged by the individual results, the 2:1 mixture of 8 was
subjected to hydrogenolysis/decarboxylation to give the desired
b-methyl product 9 in 95% yield along with 3–4% of 4a.
N-desilylation of 9 was achieved using TBAF in CH2Cl2 at 0 °C
giving 2c in 85% yield.6 Debenzylation of 2c was accomplished
under more vigorous hydrogenolysis conditions in EtOH (10%
Pd/C, 45 psi H2, 1 h) to give 2a in 90% yield.7,8
Although the malonic acid series demonstrated a ster-
eospecific decarboxylation,4d this was not an issue in the
decarboxylation of 8 (via 3c). This could be due to the greater
stability of the enol intermediate and the corresponding lower
transition state energy vs. the ketene acetal intermediate in the
malonic acid series. The greater selectivity in the protonation
step may be a result of the lower reaction temperature (20 °C for
2a vs. 80 °C for 2b) as well as the relative energetics of the two
systems.
HO
TBMSO
CH3
CH3
H
H
N
1
H
H
R
Y
NH
O
O
O
–
CO2
2a R = CH2OH
b R = OH
c R = CH2OBn
d R = OCH3
TBMSO
TBMSO
CH3
CH3
CO2H
R
H
H
N
H
R
O
O
O
TBDMS
O
NH R′
3a R = CH2OH
b R = OH
c R = CH2OBn
d R = OCH3
4a R = CH2OBn, R′ = TBDMS
b R = OCH3, R′ = TBDMS
c R = CH2OBn, R′ = H
Some time ago, we (and others) reported that the malonic acid
3b underwent a highly stereoselective decarboxylation/protona-
tion to give the 1b-methyl carboxylic acid 2b.4a,b We envi-
sioned that the keto acid 3a or 3c should undergo a similar
decarboxylation to give 2a or 2c, respectively.
After several unsuccessful attempts to generate the requisite
b-keto acid 3a via aqueous ester saponification, we adjusted our
strategy to allow preparation of a b-keto acid in a nonpolar
solvent via hydrogenolysis of a suitable benzyl ester. Benzyl
4-benzyloxy-2-methyl-3-oxobutyrate 5 was prepared by the
Claisen condensation of benzyl propionate (2 equiv. + 2 equiv.
LDA) and methyl O-benzylglycolate in 80% yield (Scheme 1).
The acetoacetate 5 was then coupled with acetoxyazetidinone 6
(K2CO3 in DMF at 45 °C, 1 h) to give ca. 2:1 diastereomeric
mixture 7 in 85–90% yield. N-silylation of 7 was carried out
using TBDMSOTf and Et3N in DMF to give 8 in quantitative
yield. In our previous work on the malonic acid series we had
The Authors thank Dr Mark Greenlee for helpful suggestions
during the preparation of this manuscript.
Notes and References
† E-mail: joe_lynch@merck.com
1 A. Tomasz, New England J. Med., 1994, 330, 1247; F. C. Tenover and
J. M. Hughes, JAMA, 1996, 275, 300.
2 (a) S. Schmitt, T. N. Salzmann, D. Shih and B. G. Christensen,
J. Antibiot., 1988, 780; (b) M. Imuta, H. Itani, H. Ona, T. Konoike, S.
Uyeo, Y. Kimura, H. Miwa, S. Matsuura and T. Yoshida, Chem. Pharm.
Chem. Commun., 1998
1817
Choi, Woo-Baeg
