A highly stereoselective and practical total synthesis of the tricyclic β-lactam antibiotic GV104326 (4-methoxytrinem)

Article abstract of DOI:10.1021/ja962006n

A novel and practical total synthesis of a tricyclic β-lactam antibiotic, GV104326 (4-methoxytrinem or Sanfetrinem) has been achieved in nine steps and about 33% overall yield from a commercially available acetoxyazetidinone chiron. A key step in the high

Full text of DOI:10.1021/ja962006n

9884
J. Am. Chem. Soc. 1996, 118, 9884-9891
A Highly Stereoselective and Practical Total Synthesis of the
Tricyclic â-Lactam Antibiotic GV104326 (4-Methoxytrinem)
Stephen Hanessian* and Michael J. Rozema†
Contribution from the Department of Chemistry, UniVersite´ de Montre´al, P. O. Box 6128,
Succ. Centre-Ville, Montre´al, QC, Canada H3C 3J7
ReceiVed June 14, 1996X
Abstract: A novel and practical total synthesis of a tricyclic â-lactam antibiotic, GV104326 (4-methoxytrinem or
Sanfetrinem) has been achieved in nine steps and about 33% overall yield from a commercially available
acetoxyazetidinone chiron. A key step in the highly diastereoselective synthesis is a protonation of a zinc enolate
complex which circumvents the use of enantiomerically pure (S)-2-methoxycyclohexanone. A mechanistic rationale
is presented and experimentally verified.
Due to the growing number of bacterial strains acquiring
resistance to the current arsenal of chemotherapeutic drugs, the
search for novel mechanism-based antibacterial agents continues
to be a major area of research on many fronts. The effort in
molecules containing the â-lactam motif alone has produced a
plethora of novel chemical entities many of which are presently
marketed as antibiotic agents.1
The report of the discovery of thienamycin by scientists at
the Merck Laboratories in 19762 opened a new frontier in
â-lactam research. Since then, great advances have been made
in the chemistry and biology of the carbapenem class of
â-lactam antibiotics. The superior stability of 1-substituted
carbapenems toward dehydropeptidases3 and the recent reports
of interesting analogs having polar substituents at C-14 are just
two examples of chemical modifications that have generated
broad spectrum antibacterial activity.
Recently, the scientists at Glaxo have discovered a new family
of synthetic â-lactam derivatives which bear the name “trinems”
(previously referred to as tribactams)5 (Figure 1). The incor-
poration of a cyclohexane ring onto the carbapenem nucleus
provides sites for chemical functionalization, as exemplified by
substituents at C-4 (Figure 1). The most promising candidate
in this class, sodium (4S,8S,9R,10S)-4-methoxy-10-[(1R)-1-
hydroxyethyl]-11-oxo-1-azatricyclo [7.2.0.03,8]undec-2-ene-2-
carboxylate (2, X ) OMe, R ) Na, GV104326), has shown
Figure 1.
excellent activity against a wide range of bacteria including
â-lactamase producing strains and is presently undergoing
clinical evaluation.6
The “trinem” 2 and its analogs present a number of synthetic
challenges not the least of which is the development of an
industrially viable route that is amenable to large scale produc-
tion. Continuing our interest in the synthesis of structurally
novel and diverse â-lactam-containing molecules,7 we have
explored methodology directed toward the synthesis of the
trinems. Paramount in the synthesis of these molecules is the
control of the configurations at C-4 and C-8 at strategic points
in the elaboration of the tricyclic nucleus. The published
syntheses of the 4-methoxy trinem (GV104326)5a,e proceed
through intermediate 4 (Figure 2), which undergoes cyclization
† Present address: Abbott Laboratories, Abbott Park, North Chicago,
IL 60064.
X Abstract published in AdVance ACS Abstracts, September 15, 1996.
(1) For recent reviews on â-lactam antibotics, see: (a) Berko, A. H.
Tetrahedron 1996, 52, 331. (b) The Organic Chemistry of â-Lactams; Georg,
G I., Ed.; VCH Publications, Inc.: New York, 1993 and references cited
therein. (c) Neuhaus, F. C.; Georgopapadakou, N. H. In Emerging Targets
in Antibacterial and Antifungal Chemotherapy; Sutcliffe, J., Georgopa-
padakou, N. H., Eds.; Chapman and Hall: New York, 1992. (d) Georgo-
papadakou, N. H. Antimicrobial Agents Ann. 1988, 3, 409.
(2) (a) Melillo, D. G.; Cretovich, R. J.; Ryan, K. M.; Sletzinger, M. J.
Org. Chem. 1986, 51, 1498. (b) Karady, S.; Amato, J. S.; Reamer, R. A.;
Weinstock, L. M. J. Am. Chem. Soc. 1981, 103, 6765. (c) Saltzmann, T.
N.; Ratcliffe, R. W.; Christensen, B. G.; Bouffard, F. A. J. Am. Chem. Soc.
1980, 102, 6161. (d) Ratcliffe, R. W.; Albers-Scho¨nberg, G. In Chemistry
and Biology of â-Lactam Antibiotics; Morin, R. B., Gorman, M., Eds.;
Academic Press: New York, 1982; Vol. 2, pp 217-314. (e) Kahan, J. S.;
Kahan, F.; Goegelman, R.; Currie, S. A.; Jackson, M.; Stapely, E. O.; Miller,
T. W.; Miller, A. K.; Hendin, D.; Mochales, S.; Hernandez, S.; Woodruff,
H. B.; Birnbaum, J. J. Antibiot. 1979, 32, 1. (f) Albers-Scho¨nberg, G.; et
al. J. Am. Chem. Soc. 1978, 100, 6491.
(5) (a) Camerini, R.; Panunzio, M.; Bonanomi, G.; Donati, D.; Perboni,
A. Tetrahedron Lett. 1996, 37, 2467. (b) Di Fabio, R.; Feriani, A.; Gaviraghi,
G.; Rossi, T. Bioorg. Med. Chem. Lett. 1995, 5, 1235. (c) Padova, A.;
Roberts, S. M.; Donati, D.; Marchioro, C.; Perboni, A. J. Chem. Soc., Chem.
Commun. 1995, 661. (d) Padova, A.; Roberts, S. M.; Donati, D.; Perboni,
A.; Rossi, T. J. Chem. Soc., Chem. Commun. 1994, 441. (e) Perboni, A.;
Tamburini, B.; Rossi, T.; Donati, D.; Tarzia, G.; Gaviraghi, G. In Recent
AdVances in the Chemistry of Anti-InfectiVe Agents; Bentley, H. H.,
Ponsford, R., Eds; The Royal Society of Chemistry, T. Graham House:
Cambridge, U.K., 1992; p 21. (f) Tamburini, B.; Perboni, A.; Rossi, T.;
Donati, D.; Andreotti, D.; Gaviraghi, G.; Carlesso, R.; Bismara, C. Eur.
Pat. Appl. EP 416 953 (cl. C07D477/00), 13 March 1991. Chem. Abstr.
1992, 116, 235 337t. Perboni, A.; Bismara, C.; Pentassuglia, G. Eur. Pat.
Appl. EP 0 466 509.
(3) Shih, D. H.; Baker, F.; Cama, L.; Christensen, B. G. Heterocycles
1984, 21, 29.
(4) (a) Mastalerz, H.; Menard, M. J. Org. Chem. 1994, 59, 3223. (b)
Rao, V. S.; Fung-Tomc, J. C.; Desiderio, J. V. J. Antibiot. 1993, 46, 167.
See also, ref 7a.
(6) Di Modugno, E.; Erbetti, I.; Ferrari, L.; Galassi, G.; Hammond, S.
H.; Xerri, L. Antimicrob. Agents Chemother. 1994, 38, 2362.
S0002-7863(96)02006-9 CCC: $12.00 © 1996 American Chemical Society

Total Synthesis of 4-Methoxytrinem
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9885
Figure 2.
to the tricyclic structure by a trialkyl phosphite-mediated
reaction.5f This intermediate is produced by elaboration of the
corresponding cyclohexene derivative. Several improvements
have been made in the original synthesis,8 which involves the
use of enantiomerically pure (S)-2-methoxycyclohexanone. The
quest for a convergent total synthesis avoiding the use of a chiral
nonracemic cyclohexanone derivative is a much sought after
objective.
leading to stereodifferention in the protonation step, resulting
in a preponderance of the desired (2S) enantiomer 4 over the
alternative (2R) isomer (cyclohexanone numbering, Figure 2).
Our synthesis started with the commercially available 2-meth-
oxycyclohexanone (Scheme 1), which was (allyloxy)carbony-
lated using diallyl carbonate and sodium hydride in the presence
of a catalytic amount of potassium hydride to produce the
desired racemic â-ketoester 7.13 Reaction of the sodium salt
of 7 with the commercially available acetoxyazetidinone 6 at
-20 °C gave a mixture of the four possible diastereomers shown
in expression 8, which were N-protected as the tert-butyldim-
ethylsilyl (TBS) derivative 9. The stereochemistry at C-8 could
now be secured by using a modified procedure of the de-
(allyloxy)carbonylation reaction.12 Thus, treatment of the four
diastereomers of 9 with 10 equiv of formic acid in the presence
of a catalytic amount of palladium acetate provided a mixture
of the N-TBS-protected compounds 10, 11, and 12 (ca. 4:4:1).
At this point, it was possible to separate 10 from 11 and 12 for
characterization. Selective N-deprotection of 10 provided 4,
while using the same conditions, 11 and 12 could be selectively
deprotected to provide a mixture of 13 and 14 which could be
separated. Routinely on larger scales, the first four steps were
conducted with only one chromatographic separation at the end
of a sequence to remove 14, leaving a mixture of 4 and 13 (ca.
1:1 at C-4).
At this point we were faced with the task of finding the proper
conditions for the conversion of the (2R) isomer 13 in the
mixture of methoxy epimers to the desired (2S)-methoxy ketone
intermediate 4. Treatment of the â-methoxy derivative 13 with
lithium diisopropylamide (LDA, 3 equiv, -78 °C, 30 min)
followed by quenching the reaction with D2O gave >95%
incorporation of deuterium, showing that the formation of the
dianion was possible. However, the products consisted of a
1:1.5 mixture of R- and â-methoxy derivatives 4 and 13,
respectively (Table 1, entry 1). The literature contains some
reports of stereoselective protonations of enolates involving a
variety of other proton sources.14 Thus, methyl (S)-lactate
(Table 1, entry 2), diethyl malonate (Table 1, entry 3), and
pyrrolidinone (Table 1, entry 5) were tried as proton sources,
but no stereochemical preference was seen in the mixture of
epimers.
Results and Discussion
We wish to report a practical and highly convergent total
synthesis of the 4-methoxytrinem (GV104326), utilizing a
strategy in which the stereochemical control in a key step is
derived entirely from the inherent chirality of the commercially
available azetidinone 6.9 Our disconnective analysis of the
target structure is shown in Figure 2, and as in the previous
syntheses, it hinges on the preparation of 4 in enantiomerically
pure form. In our recently reported synthesis of the 5-meth-
oxytrinem,10 we showed that the coupling of 3-methoxy-6-
((allyloxy)carbonyl)-2-cyclohexen-1-one with the azetidinone
6,11 followed by protection of the lactam and subsequent de-
(allyloxy)carbonylation12 provided a highly stereocontrolled
access to an intermediate in which the desired â-configuration
at the cyclohexanone substitution site was secured.
Our strategy for the synthesis of 2 also relied on a similar
coupling protocol. In addition, we also wished to avoid the
use of enantiomerically pure cyclohexanone derivatives. We
hoped that given the chiral nonracemic nature of the intended
substrate 4, we would be able to adjust the configuration at the
methoxy center through an enolate such as 5. We reasoned
that with the appropriate choice of the metal cation (chelating
or nonchelating) we could fix the conformation of the enolate
(7) (a) Hanessian, S.; Rozema, M. J. Bioorg. Med. Chem. Lett. 1994, 4,
2279. (b) Hanessian, S.; Reddy, G. B. Bioorg. Med. Chem. Lett. 1994, 4,
2285. (c) Hanessian, S.; Couture, C. A.; Georgopapadakou, N. Bioorg. Med.
Chem. Lett. 1993, 3, 2323. (d) Sumi, K.; Di Fabio, R.; Hanessian, S.
Tetrahedron Lett. 1992, 33, 749. (e) Hanessian, S.; Sumi, K.; Vanasse, B.
Synlett 1992, 33. (f) Hanessian, S.; Desilets, D.; Bennani, Y. L. J. Org.
Chem. 1990, 55, 3098. (g) Hanessian, S.; Desilets, D. In Trends in Medicinal
Chemistry; van der Groot, H., Domany, G., Pallos, L., Timmerman, H.,
Eds.; Elsevier: Amsterdam, 1989; p 165. (h) Hanessian, S.; Alpegiani, M.
Tetrahedron 1989, 45, 941. (i) Hanessian, S.; Bedeschi, A.; Battistini, C.;
Mongelli, N. J. Am. Chem. Soc. 1985, 107, 1438. See also, ref 10.
(8) (a) Ghiron, C.; Piga, E.; Rossi, T.; Tamburini, B.; Thomas, R. J.
Tetrahedron Lett. 1996, 37, 3891. (b) Rossi, T.; Biondi, S.; Contini, S.;
Thomas, R. J.; Marchioro, C. J. Am. Chem. Soc. 1995, 117, 9604. (c)
Bismura, C.; Di Fabio, R.; Donati, D.; Rossi, T.; Thomas, R. J. Tetrahedron
Lett. 1995, 36, 4283. See also, Rossi, T.; Zarantonello, P.; Thomas, J. R.
PCT Appl. WO95/26333.
Given these results, we decided to study the protonation of
the dianion as expressed in 5 in the presence of a bidentate
chelating metal. We reasoned that, in addition to a favorable
geometry, the metal would provide a coordination site onto
which a suitable reagent could anchor itself for a more selective
(9) 4-Acetoxy-3-[(R)-1-((tert-butyldimethylsilyl)oxy)ethyl]-2-azetidino-
ne is commercially available from Kaneka America Corp., New York, NY,
or Aldrich Chemical Co., Milwaukee, WI.
(13) Greengrass, C. W.; Hoople, D. W. T. Tetrahedron Lett. 1981, 22,
1161.
(10) Hanessian, S.; Rozema, M. J.; Reddy, G. B.; Braganza, J. F. Bioorg.
Med. Chem. Lett. 1995, 5, 2535.
(11) Paquette, L. A.; Dahnke, K.; Doyon, J.; He, W.; Wyant, K.;
Friedrich, D. J. Org. Chem. 1991, 56, 6199.
(12) (a) Murayama, T.; Yoshida, A.; Kobayashi, T.; Miura, T. Tetrahe-
dron Lett. 1994, 35, 2271. (b) Choi, W.-B.; Churchill, H. R. O.; Lynch, J.
E.; Thompson, A. S.; Humphrey, G. R.; Volante, R. P.; Reider, P. J.; Shinkai,
I. Tetrahedron Lett. 1994, 35, 2275. (c) Jones, D. K.; Liotta, D. C.; Choi,
W.-B.; Volante, R. P.; Reider, P. J.; Shinkai, I.; Churchill, H. R. O.; Lynch,
J. E. J. Org. Chem. 1994, 59, 3749. (d) Shimamoto, T.; Inoue, H.; Yoshida,
T.; Tanaka, R.; Nakatsuka, T.; Ishiguro, M. Tetrahedron Lett. 1994, 35,
5887.
(14) (a) Fuji, K.; Kawabata, T.; Kuroda, A. J. Org Chem. 1995, 60, 1914.
(b) Fehr, C.; Galindo, J. Angew. Chem., Int. Ed. Engl. 1994, 33, 1888. (c)
Krause, N. Angew. Chem., Int. Ed. Engl. 1994, 33, 1764. (d) Gerlach, U.;
Haubenreich, T.; Hu¨nig, S. Chem. Ber. 1994, 127, 1969. (e) Gerlach, U.;
Haubenreich, T.; Hu¨nig, S. Chem. Ber. 1994, 127, 1981. (f) Gerlach, U.;
Haubenreich, T.; Hu¨nig, S.; Klaunzer, N. Chem. Ber. 1994, 127, 1989. (g)
Baker, W.; Pratt, J. K. Tetrahedron 1993, 49, 8739. (h) Matsumoto, K.;
Ohta, H. Tetrahedron Lett. 1991, 32, 4729. (i) Larsen, R. D.; Corley, E.
G.; Davis, P.; Reider, P. J.; Grabowski, E. J. J. J. Am. Chem. Soc. 1989,
111, 7650. (j) Gerlach, U.; Hu¨nig, S. Tetrahedron Lett. 1987, 28, 5805. (k)
Duhamel, L.; Fouquay, S.; Plaquevent, J.-C. Tetrahedron Lett. 1986, 27,
4975.

9886 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Hanessian and Rozema
Scheme 1
Table 1. Protonation of the Lithium Enolate of 5
Table 2. Protonation of the Zinc Enolate of 5
entry
proton source
R/â ratioa
entry
proton source
R/â ratioa
1
2
3
4
5
D2O
1:1.5b
1.2:1
1:1
1:1.2
1:1.5
1
2
3
4
5
6
7
8
D2O
pyrrolidinone
diethyl malonate
malononitrile
ethyl cyanoacetate
nitromethane
2-pyridylacetonitrile
2-mercaptopyridine
Meldrum’s acid
2.5:1
1.5:1
>15:1
7:1
methyl (S)-lactate
(EtO2C)2CH2
TMEDA:(EtO2C)2CH2
pyrrolidinone
>15:1
2:1
a Determined by integrating the methoxy CH3 resonances (R-
methoxy, 3.29 ppm; â-methoxy, 3.46 ppm) in 1H NMR (300 MHz)
spectrum of crude reaction mixture. b D incorporation >95%.
5:1c
1:2
3:1
9
10
11
12
13
diethyl methylmalonate
ethyl acetoacetate
di-tert-butyl malonate
bromozinc diethyl malonateb
>15:1
1:1.5c
9:1
proton delivery. Formation of the lithium dianion followed by
addition of ZnBr2 (1.1 equiv) gave, after quenching the reaction
with D2O, a 2.5:1 mixture of 4 and 13, respectively. Curiously,
no deuterium was incorporated into the mixture of epimers,
indicating the likehood of an internal proton quench. A series
of experiments were then carried out in order to further improve
the stereoselectivity in favor of 4. We were thus pleased to
find that addition of diethyl malonate (5 equiv) to the zinc
dianion followed by aqueous workup gave the R- and â-methoxy
isomers 4 and 13 in a ratio of >15:1 respectively, and that the
R-methoxy isomer 4 could be isolated in a 77% yield. When
subjected to the same conditions of dianion formation and
protonation, the R-methoxy isomer 4 was recovered unchanged
in 73% yield. Treatment of the â-methoxy isomer 13 under
the same conditions resulted in a nearly total epimerization to
4 (72% yield). On a preparative scale, a mixture of methoxy
isomers 4 and 13 isolated after condensation with the azetidinone
6 could be transformed to the desired R-methoxy isomer 4 in
over 72% yield, with a major portion being obtained by direct
crystallization before chromatography. A remarkably efficient
and stereoselective route to the intended critical intermediate 4
was thus in hand.
3:1
a Determined by integrating the methoxy CH3 resonances (R-
1
methoxy, 3.29 ppm; â-methoxy, 3.46 ppm) in H NMR (300 MHz)
spectrum of crude reaction mixture. b Prepared by deprotonating diethyl
malonate with LDA and adding zinc bromide (1 equiv). c Significant
amounts of unidentified side products were also produced.
source of the proton were systematically studied. To prove the
existence of the chelation to the zinc, we conducted two crucial
experiments. Protection of the lactam nitrogen with TBS triflate
and formation of the zinc enolate (LDA, 1.1 equiv) followed
by addition of ZnBr2 (1 equiv) and quenching the reaction with
diethyl malonate resulted in a complete reversal of selectivity
in protonation (<1:15, R/â). Also, all selectivity could be
obliterated by performing the reaction in the presence of 10
equiv of zinc bromide. These experiments strongly suggested
the presence of an intramolecular chelation of the enolate oxygen
and the lactam nitrogen to the zinc.
In order to address the role of the diethyl malonate, a set of
potential proton sources were selected and added to the zinc
dianion of 4 and/or 13 given as 5. The results are listed in
Table 2.
With the exception of 2-mercaptopyridine (Table 2, entry 8)
and ethyl acetoacetate (Table 2, entry 11), the protonation of
To understand the nature of this highly selective protonation,
a series of experiments were designed where the role of the
zinc bromide, the role of the diethyl malonate, and the exact

Total Synthesis of 4-Methoxytrinem
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9887
Table 3. Deuteration of the Zinc Enolate of 5
entry
base
additive
(EtO2C)2CD2
(EtO2C)2CH2
(EtO2C)2CD2
(EtO2C)2CD2
(EtO2C)2CD2
(EtO2C)2CD2
(EtO2C)2CD2
(EtO2C)2CD2
(EtO2C)2CD2
(EtO2C)2CMe2
i-Pr2ND; (EtO2C)2CMe2
i-Pr2ND; ZnBr2; (EtO2C)2CMe2
(EtO2C)2CMe2 (3 h); i-Pr2ND
time (h)
workup
R/âa
% Db
1
2
3
4
5
6
7
8
LDA (3 equiv)
LDA (3 equiv)
LDA (3 equiv)
LDA (3 equiv)
LDA (3 equiv)
LTMP (3 equiv)
LiHMDS (3 equiv)
LiHMDS (3 equiv)
LDA (1.9 equiv)
LDA (3 equiv)
LDA (3 equiv)
LDA (3 equiv)
LDA (3 equiv)
10c
10c
6
6
6
6
6
6
6
6
6
6
3
NH4Cl
D2O
NH4Cl
D2O
NH4Cl
NH4Cl
NH4Cl
D2O
NH4Cl
D2O
>15:1
>15:1
>15:1
>15:1
>15:1
>15:1
>15:1
>15:1
>15:1
>15:1
>15:1
>15:1
>15:1
0
0
65
65
50
50
65
65
0
30
50
50
20
9
10
11
12
13
NH4Cl
NH4Cl
NH4Cl
1
a Determined by integrating the methoxy CH3 resonances (R-methoxy, 3.29 ppm; â-methoxy, 3.46 ppm) in H NMR (300 MHz) spectrum of
1
crude reaction mixture. b Determined by integrating the disappearance of the resonance at 3.58 ppm; (H R to OMe) in the H NMR (300 MHz)
spectrum of crude reaction mixture. c Time is given in minutes.
product. This result suggested that the proton was actually being
delivered by a source other than the malonate, for example
during the aqueous workup or by the diisopropylamine generated
upon formation of the lithium enolate. The zinc enolate was
therefore treated with diethyl malonate, and subsequently the
reaction was quenched with D2O. Except for the excellent
selectivity the product was once again devoid of deuterium
(Table 3, entry 2), similar to when ZnBr2 alone had been used
(vide supra). The only other source of proton that remained
was the diisopropylamine generated during the formation of the
enolate, which could deliver a proton internally.16
Figure 3.
the zinc enolate 5 leads preferentially to the formation of the
R-methoxy derivative 4. It is believed that the 2-mercaptopy-
ridine and the ethyl acetoacetate are too strongly coordinating,
which results in breaking the chelation of the zinc to the enolate
oxygen and the lactam nitrogen. A weak preference for the
â-methoxy epimer 13 over the desired 4 was in fact observed.
The remaining entries in Table 2 show that quenching the
zinc enolate with a variety of other proton sources leads to a
preferential formation of the R-methoxy epimer 4. The highest
diastereoselection was observed with diethyl malonate, ethyl
cyanoacetate, and diethyl methylmalonate (Table 2, entries 3,
5, and 10, respectively). Other carbon acids such as malono-
nitrile, 2-pyridylacetonitrile, and di-tert-butyl malonate (Table
2, entries 4, 7, and 12, respectively) showed modest to fairly
good selectivities. Quenching with pyrrolidinone, nitromethane,
and Meldrum’s acid and using the bromozinc salt of diethyl
malonate (Table 2, entries 2, 6, 9, and 12, respectively) gave
poor selectivities, albeit always in favor of the R-methoxy isomer
4. From this data, we hypothesized the existence of an
intermediate where malonate could occupy the axial ligand
position on the zinc enolate as shown in Figure 3. This would
situate the methylene protons of the malonate moiety for
optimum delivery to the â-face of the C-4 enolate atom. An
equatorial disposition of the malonate places the methylene
protons in a geometrically unfavorable position relative to the
enolate R-carbon. The presence of excess malonate in solution
would ensure the presence of the neutral molecule as a ligand,
rather than as its enolate.
Other researchers have found that the addition of butyllithium
alleviates the problem of incomplete deuteration by metalating
the liberated diisopropylamine.15 Thus, addition of butyllithium
to the lithium enolate, followed by the addition of zinc bromide
and diethyl malonate-d2 and aqueous NH4Cl workup after 6 h
of stirring at room temperature, provided the R-methoxy
derivative 4, in which 65% deuterium was introduced (Table
3, entry 3). A longer contact time with the deuterium source
(malonate-d2 or diisopropylamine-d) was required to reach this
degree of deuteration. The same level of deuterium incorpora-
tion was also achieved when D2O was used in the workup (Table
3, entry 4). When the conditions used in entry 1 were repeated
with a longer reaction time, a 50% D incorporation was obtained
(Table 3, entry 5). Reasonably good levels of deuterium
incorporation could also be obtained with lithium tetrameth-
ylpiperidine (LTMP, Table 3, entry 6) and lithium hexameth-
yldisilazane (LiHMDS, Table 3, entries 7 and 8). However,
when a substoichiometric quantity of LDA (Table 3, entry 9)
was used in conjunction with deuterated malonate, the level of
deuterium incorporation fell to zero while a high level of
selectivity was maintained (Table 3, entry 9). Although these
results show that deuteration is possible under specific conditions
of stoichiometry, reaction time, and workup, they do not identify
the exact source of the deuteron or, for that matter, the proton
donor. A distinct possibility emerges where addition of diethyl
malonate-d2 to the solution which contains LDA produces
deuterated diisopropylamine, which could be the enigmatic
deuteron donor.
We now had to validate the proposed internal delivery of
protons by the zinc-anchored malonate as shown in Figure 3.
An obvious way to answer to this question was to perform a
series of deuteration experiments, the results of which are shown
in Table 3.
Diethyl malonate-d2 was added to the zinc enolate 5 (Figure
2), the solution was stirred at room temperature for 10 min,
and then the reaction was quenched with aqueous ammonium
chloride (Table 3, entry 1). While examples of incomplete
deuteration of enolates are known in the literature,15 we were
puzzled by the fact that no deuterium was incorporated in the
The longer reaction time in the presence of deuterated
malonate and the remaining LDA (1 equiv) allows for deuterium
exchange to take place with the diisopropylamine in solution.
Evidently, a shorter reaction time does not allow for deuterated
diisopropylamine to exchange with the initially coordinated
diisopropylamine in the zinc enolate complex. In fact, deute-
(15) For an excellent review on the structure and reactivity of lithium
enolates, see: Seebach. D. Angew. Chem., Int. Ed. Engl. 1988, 27, 1624.
(16) Vedejs, E.; Lee, N. J. Am. Chem. Soc. 1991, 113, 5483.

9888 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Hanessian and Rozema
rium exchange between malonate-d2 and diisopropylamine was
only partially complete after 18 h at 25 °C, as evidenced by 1H
NMR. Although the exchange is expected to be instantaneous
with the additional equivalent of LDA still present in the reaction
mixture (Table 3, entries 1 and 2), the presence of excess
malonate-d2 (4 equiv) and the slower exchange of unbound
diisopropylamine-d for the bound amine accounts for the
nonincorporation of deuterium in entry 1. The result in entry
2 would argue in favor of a rapid internal proton transfer from
bound diisopropylamine before the workup. The possible
exchange of deuterium from D2O for hydrogen on the coordi-
nated diisopropylamine during the workup may be too slow to
compete with internal delivery.
Figure 4.
Table 4. Quenching of the Zinc Enolate of 5 with Nonprotic
Additives
The decisive experiments were then to remove the protons
from the malonate and to allow the diisopropylamine to be the
only source of proton prior to the workup. Entries 10-13 of
Table 3 show the results of these experiments. Replacing the
diethyl malonate with diethyl 2,2-dimethylmalonate and quench-
ing the reaction with D2O produces the R-methoxy derivative
4 with the same high leVel of diastereoselectiVity (>15/1) as
before but with only 30% D incorporation (Table 3, entry 10).
We then conducted the reaction in the presence of excess
diisopropylamine-d prior to and after the addition of zinc
bromide to the enolate (Table 3, entries 11 and 12). Again, a
50% D incorporation was observed with high selectivity in both
cases, thus demonstrating that the order of addition of reagents
to the initial enolate was not critical, as long as the zinc bromide
and malonate were present in conjunction with diisopropy-
lamine-d. It was thus clear that the only source of deuterium
in these reactions was the diisopropylamine-d and that a contact
time of at least 6 h was necessary for reasonable to good levels
of deuterium incorporation. We should recall that a D2O quench
of the lithium enolate 5 (Figure 2) gave a mixture of 4 and the
C-2 epimer 11, in which >95% deuterium had been incorpo-
rated. Since the only difference is the presence of the malonate
and the zinc enolate complex, the incomplete incorporation of
deuterium in some of the experiments described above, even
upon addition of D2O (or 4 equiv of AcOH-d4), must reflect
upon the existence of a highly organized (and possibly ag-
gregated) zinc enolate complex.
entry
additive
R/âa
1
2
3
4
diethyl dimethylmalonate
benzaldehyde
benzophenone
>15:1
>15:1
6:1
tert-butyl acetate
5:1
a Determined by integrating the methoxy CH3 resonances (R-
1
methoxy, 3.29 ppm; â-methoxy, 3.46 ppm) in H NMR (300 MHz)
spectrum of crude reaction mixture.
amine moiety is favorably deployed to deliver a proton (or
deuteron) to the â-face of the enolate leading to the desired
product. To the extent that deuterium incorporation occurs in
spite of a protic workup, we suggest a directed internal transfer
from the coordinated amine. Although the level of deuterium
incorporation increased with longer contact time (6 h vs 10 min
or 1 h) before aqueous workup, reaching as high as 65% (Table
3, entries 7 and 8), we could not extend the reaction times longer
at room temperature due to unwanted side reactions.
It is highly probable that, upon addition of water (or D2O),
there is an initial exchange of protons (or deuterons) on the
coordinated axially disposed diisopropylamine ligand before the
chelate breaks. The transfer may still take place internally, but
the level of deuterium incorporation will depend to what extent
the diisopropylamine is fully deuterated. An alternative proposal
is that a significant amount of proton transfer must occur from
the intermediate in Figure 4 prior to the workup. When
deuterium transfer is involved, there must be enough contact
time to allow the exchange to occur with bound and unbound
diisopropylamine. Unfortunately, the exact timing of proton
transfer cannot be ascertained because of the complexity of the
exchange process, the lack of definitive structural information
on the zinc enolate complex, and its stability in the reaction
medium or upon workup. Regardless of the exact nature of
the proton (or deuteron) transfer, the process is highly selective
and must involve a geometrically well-organized complex which
allows for a directed protonation.
We were intrigued by the fact that protons on the malonate
were not required, and we briefly looked into other types of
“carbon acids” (Table 4). Addition of benzaldehyde to the zinc
enolate also gives a high selectivity of protonated R-methoxy
ketone 4 with no trace of an aldol product (Table 4, entry 2).
The addition of benzophenone (Table 4, entry 3) and tert-butyl
acetate (Table 4, entry 4) also led to a relatively high degree of
selectivity toward the R-methoxy ketone 4. We assume that
the role of the diisopropylamine as an internal proton source
remains the same as that for malonates.
Having devised a highly selective and practical synthesis of
the key intermediate 4, we proceeded with the synthesis of the
final target using standard protocols. Acylation of the lactam
nitrogen with benzyl oxalyl choride produced the oxalimide 15
in 97% yield, which was subjected to triethyl phosphite-mediated
ring closure17 to provide the protected trinem 16 in 83% yield.
Desilylation with tetrabutylammonium fluoride buffered with
Entry 13 in Table 3 demonstrates the effect of contact time
with diisopropylamine-d on deuterium incorporation. The
reaction was conducted with diethyl 2,2-dimethylmalonate as
the additive for 3 h, after which excess diisopropylamine-d was
added and the reaction was continued for another 3 h. After
the reaction was quenched with NH4Cl, the deuterium content
of the highly enriched 4:13 mixture was 20%. Thus, it is
possible that a significant portion of protons had been transferred
during the first 3 h period from internally coordinated diiso-
propylamine and that exchange with diisopropylamine-d was
not efficient enough to produce more than 20% incorporation.
Had the aqueous medium of the workup been solely responsible
for the quenching of the enolate, no deuterium incorporation
should have been observed. Furthermore, the origin of the
stereochemical preference would still remain unresolved.
The questions that still remained were concerned with the
spatial disposition of the diisopropylamine relative to the zinc
enolate, the origin of the “directed” protonation and possibly
the “timing” of such an event. Data from various experiments
in Table 3 lead us to conclude that the malonate and diisopro-
pylamine entities must adopt preferred coordination sites in the
zinc enolate complex. We therefore propose a strongly
coordinated “organized” intermediate, as shown in Figure 4, in
which the axial coordination site on the zinc is occupied by the
diisopropylamine and the equatorial site by the malonate. The

Total Synthesis of 4-Methoxytrinem
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9889
butyldimethylsilyl trifluoromethanesulfonate (2.6 mL, 11.3 mmol) were
added. After 30 min of stirring, the reaction was poured into water
(50 mL) and the aqueous layer was extracted three times with CH2Cl2
(100 mL). The combined organic layers were dried over MgSO4,
filtered, and concentrated to give the crude N-TBS-protected products
9 as a pale yellow oil.
acetic acid followed by hydrogenolysis in the presence of the
bulky amidine 1818 afforded the final product 19 as a white
amorphous solid. This was converted to the corresponding
potassium salt using potassium ethylhexanoate, whose spectro-
scopic properties were in full agreement with those of an original
sample data provided to us by Glaxo (Verona, Italy).
Palladium acetate (17 mg, 0.075 mmol, 1 mol %) and triphenylphos-
phine (100 mg, 0.38 mmol, 5 mol %) were suspended in ethyl acetate
(15 mL), and formic acid (2.8 mL, 75 mmol) was added. After this
mixture was refluxed for 30 min, a solution of the crude N-TBS-
protected products 9 in ethyl acetate (35 mL) was added dropwise over
30 min. The mixture was heated at reflux for 2.5 h, cooled, diluted
with ethyl acetate (200 mL), and washed with saturated aqueous
NaHCO3 (50 mL) and brine (50 mL). After the organic phase was
dried over MgSO4, it was filtered through a layer of silica gel.
Concentration and flash chromatography (hexanes-ethyl acetate, 10:1
to 2:1) afforded 10 (1.32g, 37%) as a clear colorless oil and a 4:1
mixture of 11 and 12 (1.53 g, 43%). For 10: [R]D -2.5° (c 1.59,
CHCl3); IR (CHCl3) 2950, 2870, 1740 (â-lactam CO), 1720 (cyclo-
hexane (cHex) CO), 1520, 1110 cm-1; MS (FAB, NBA), m/e (rel
intensity) 470.5 (MH+, 54), 454.4 (15), 412.4 (100), 185.2 (54), 153.2
(72); HRMS calcd for C24H48O4NSi2 470.3122, found 470.3104.
In conclusion, we have described a practical and novel nine-
step synthesis of the Glaxo 4-methoxytrinem GV104326 in ca.
33% overall yield from a commercially available chiron. The
highlight of the synthesis is a diastereoselective protonation of
a zinc enolate complex which circumvents the use of an
enantiomerically pure (S)-2-methoxycyclohexanone. Operation-
ally, the synthesis is amenable to scaleup and it could be
potentially applicable to a manufacturing protocol.
Experimental Section
General Procedure. 1H and 13C NMR spectra were recorded on a
Bruker 300 instrument using deuteriochloroform as solvent (CHCl3
standard, δ ) 7.26 ppm) (s, singlet; d, doublet; t, triplet; m, multiplet,
br, broad). 1H and 13C NMR data are available upon request. Infrared
spectra were recorded on a Perkin-Elmer 781 infrared spectrophotom-
eter. Optical rotations were measured on a Perkin-Elmer 241 pola-
rimeter. Combustion analyses were performed by Guelph Laboratories
Ltd., Guelph, Ontario, Canada. Column chromatography was per-
formed using the flash method.19 Melting points are uncorrected.
Allyl 3-Methoxy-2-oxocyclohexanecarboxylate (7). A slurry of
60% NaH in mineral oil (2.4 g) and 30% KH in mineral oil (∼50 mg)
was washed three times with hexanes (5 mL) and once with THF (5
mL) and then suspended in THF (20 mL) and diallyl carbonate (4.4
mL, 40 mmol). This mixture was then brought to reflux, and a solution
of 2-methoxycyclohexanone (2.5 mL, 20 mmol) in THF (5 mL) was
added dropwise over 15 min. After 12 h of reflux, the reaction mixture
was cooled to room temperature, poured into a saturated aqueous
solution of NaHCO3 (100 mL) containing concentrated NH4OH (2 mL),
and stirred at room temperature for 30 min. The aqueous layer was
acidified to pH 2 with concentrated HCl and extracted three times with
ether (100 mL). The combined organic layers were washed three times
with saturated NaHCO3 (50 mL), once with water (50 mL), and once
with brine (50 mL), and dried over Na2SO4. After filtration and
concentration, the crude oil was purified by column chromatography
(hexanes-ethyl acetate, 4:1), to afford the desired â-ketoester 7 as a
mixture of tautomers and diastereomers (2.21 g, 52%).
(3S,4R)-1-tert-Butyldimethylsilyl-3-[(1R)-1-((tert-butyldimethyl-
silyl)oxy)ethyl]-4-[(1R,3S)-3-methoxy-2-oxocyclohexyl]azetidin-2-
one (10). A slurry of 60% sodium hydride in mineral oil (540 mg,
13.5 mmol) was washed three times with hexane and suspended in
THF (25 mL). This mixture was cooled to -20 °C and a solution of
7 (1.59 g, 7.5 mmol) in THF (10 mL) was added dropwise over 30
min. After 45 min of additional stirring at -20 °C, a solution of
(3S,4R)-4-acetoxy-3-[(1R)-1-((tert-butyldimethylsilyl)oxy)ethyl]-azeti-
din-2-one (6, 2.15 g, 7.5 mmol) in THF (30 mL) was added dropwise
at -20 °C over 30 min. After an additional 2 h at -20 °C, the reaction
was quenched by the addition of acetic acid (2 mL), and the usual
workup (saturated aqueous NaHCO3, ethyl acetate) afforded the crude
products 8 as a clear oil.
(3S,4R)-3-[(1R)-1-((tert-Butyldimethylsilyl)oxy)ethyl]-4-[(1R,3S)-
3-methoxy-2-oxocyclohexyl]azetidin-2-one (4). To a solution of 10
(912 mg, 1.94 mmol) in THF (10 mL) was added acetic acid (230 µL,
4.0 mmol) and a solution of tetrabutylammonium fluoride (2.0 mL,
2.0 mmol, 1 M in THF). The reaction was stirred for 30 min. and
diluted with ethyl acetate (300 mL). The organic layer was washed
with saturated aqueous NaHCO3 (50 mL), water (50 mL), 10% aqueous
HCl (50 mL), water (50 mL), and brine (50 mL) and dried over MgSO4.
After filtration and removal of the solvents, the crude product was
purified by flash chromatography (hexanes-ethyl acetate, 4:1 to 2:1)
to afford the desired product 4 (619 mg, 90%) as a white crystalline
solid: mp 127-129 °C (hexanes); [R]D 31.9° (c 1.06, CHCl3); IR
(CHCl3) 3440 (NH), 2950, 2870, 1770 (â-lactam CO), 1740 (cHex CO),
1
1520, 1100 cm-1; H NMR (300 MHz, CDCl3) δ 5.76 (s, 1H, NH),
4.24-4.15 (m, 1H, TBSOCH), 4.02-4.00 (m, 1H, â-lactam R to N),
3.58 (apparent t, 1H, MeOCH), 3.29 (s, 3H, CH3O), 3.10 (ddd, 1H, J
) 3.8, 5.4, 12.4 Hz, HC1 cHex), 2.89 (dd, 1H, J ) 2.5, 5.7 Hz, â-lactam
R to CO), 2.28-2.21 (m, 1H, cHex), 2.15-1.96 (m, 2H, cHex), 1.73-
1.52 (m, 3H, cHex), 1.26 (d, 3H, J ) 6.2 Hz, CH3CHOTBS), 0.88 (s,
9H, t-Bu), 0.09 (s, 3H, CH3Si), 0.08 (s, 3H, CH3Si); 13C NMR (75.5
MHz, CDCl3) δ 212.8, 168.6, 84.0, 65.9, 61.0, 56.9, 49.4, 48.7, 33.6,
28.4, 25.7, 22.4, 19.0, 17.8, -4.3, -5.1; MS (FAB, NBA), m/e (rel
intensity) 356.3 (MH+, 20), 298.2 (100), 181.2 (35), 156.2 (66); HRMS
calcd for C18H34O4NSi 356.2257, found 356.2248. Anal. Calcd for
C18H34O4NSi: C, 60.81; H, 9.35; N, 3.94. Found: C, 60.60; H, 9.52;
N, 3.85.
(3S,4R)-3-[(1R)-1-((tert-Butyldimethylsilyl)oxy)silyl)oxy)ethyl]-4-
[(1R,3R)-3-methoxy-2-oxocyclohexyl]azetidin-2-one (13) and (3S,4R)-
3-[(1R)-1-((tert-Butyldimethylsilyl)oxy)ethyl]-4-[(1S,3S)-3-methoxy-
2-oxocyclohexyl]azetidin-2-one (14). To a solution of the mixture of
11 and 12 obtained above (1.25 g, 2.67 mmol) in THF (10 mL) were
added acetic acid (300 µL, 5.2 mmol) and a solution of tetrabutylam-
monium fluoride (2.8 mL, 2.8 mmol, 1 M in THF). The reaction was
stirred for 1 h and diluted with ethyl acetate (300 mL). The organic
layer was washed with saturated aqueous NaHCO3 (50 mL), water (50
mL), 10% aqueous HCl (50 mL), water (50 mL) and brine (50 mL)
and dried over MgSO4. After filtration and removal of the solvents,
the crude products were purified by flash chromatography (hexane-
ethyl acetate, 4:1 to 1:2) to afford 13 (597 mg, 63%) and 14 (128 mg,
14%) as white crystalline solids. For 13: mp 129-131 °C (hexane);
[R]D 83.4° (c 1.37, CHCl3); IR (CHCl3) 3470 (NH), 2990, 2930, 1770
(â-lactam CO), 1740 (cHex CO), 1470, 1140 cm-1; 1H NMR (300 MHz,
CDCl3) δ 5.76 (s, 1H, NH), 4.23-4.12 (m, 1H, TBSOCH), 4.09
(apparent t, 1H, â-lactam R to N), 3.81 (dd, 1H, J ) 6.4, 12.1 Hz,
MeOCH), 3.46 (s, 3H, CH3O), 2.87 (dd, 1H, J ) 2.5, 5.2 Hz, â-lactam
R to CO), 2.62-2.54 (m, 1H), 2.47-2.40 (m, 1H), 2.16-1.99 (m, 2H,
cHex), 1.81-1.57 (m, 3H, cHex), 1.25 (d, 3H, J ) 6.2Hz, CH3-
CHOTBS), 0.88 (s, 9H, t-Bu), 0.09 (s, 3H, CH3Si), 0.07 (s, 3H, CH3-
Si); 13C NMR (75.5MHz, CDCl3) δ 208.5, 168.5, 84.2, 65.6, 61.1, 57.6,
51.8, 48.7, 34.4, 28.2, 25.6, 22.4, 22.2, 17.7, -4.5, -5.2; MS (FAB,
A solution of the crude products 8 prepared above in CH2Cl2 (20
mL) was cooled to 0 °C, 2,6-lutidine (1.7 mL, 15.0 mmol) and tert-
(17) (a) Yoshida, A.; Tajima, Y.; Takeda, N.; Oida, S. Tetrahedron Lett.
1984, 25, 2793. (b) Battistini, C.; Scarafile, C.; Foglio, M.; Franceschi, G.
Tetrahedron Lett. 1984, 22, 2395. (c) Perrone, E.; Alpegiani, M.; Bedeschi,
A.; Giudici, F.; Franceschi, G. Tetrahedron Lett. 1984, 22, 2399. (d) Afonso,
A.; Hon, F.; Weinstein, J.; Ganguly, A. K.; McPhail, A. T. J. Am. Chem.
Soc. 1982, 104, 6138. (e) Ernest, I.; Gosteli, J.; Greengrass, C. W.; Holick,
W.; Jackman, D. E.; Pfaendler, H. R.; Woodward, R. B. J. Am. Chem. Soc.
1978, 100, 8214. (d) Afonso, A.; Hon, F.; Weinstein, J.; Ganguly, A. K.;
McPhail, A. T. J. Am Chem. Soc. 1982, 104, 8214.
(18) For the use of amidines as bases in salt formation, see: Shibuya,
M.; Kuretani, M.; Kubota, S. Tetrahedron 1982, 38, 2659. Heinzer, F.;
Soukup, M.; Eschenmoser, A. HelV. Chim. Acta 1978, 61, 2851. See also,
refs 7a, 7b, and 10.
(19) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923.

9890 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Hanessian and Rozema
THIO), m/e (rel intensity) 356.3 (MH+, 12), 298.5 (35), 181.1 (50),
156.2 (100), 114.9 (31); HRMS calculated for C18H34O4NSi 356.2257,
found 356.2248. For 14: mp 124-126 °C (hexane); [R]D -17.5° (c
1.18, CHCl3); IR (CHCl3) 3470 (NH), 2980, 2930, 1760 (â-lactam CO),
and either saturated aqueous ammonium chloride or D2O as specified.
Integrating the disappearance of the resonance at 3.58 ppm (H R to
1
OMe) in the H NMR spectra (300 MHz) of crude reaction mixtures
gave the values in Table 3. In entries 3 and 4, the butyllithium was
added to the lithium enolate at -78 °C and stirred for an additional
hour before the addition of the zinc bromide.
1
1720 (cHex CO), 1470, 1140 cm-1; H NMR (300 MHz, CDCl3) δ
6.07 (s, 1H, NH), 4.19-4.08 (m, 1H, TBSOCH), 3.82-3.76 (m, 1H,
ΜeOCH), 3.69 (dd, 1H, J ) 2.1, 9.7 Hz, â-lactam R to N), 3.46 (s,
3H, CH3O), 2.71 (dd, 1H, J ) 1.4, 5.7 Hz, â-lactam R to CO), 2.47-
2.35 (m, 2H), 2.17-2.10 (m, 1H), 2.02-1.96 (m, 1H), 1.80-1.57 (m,
2H), 1.44-1.27 (m, 1H), 1.24 (d, 3H, J ) 6.2 Hz, CH3CHOTBS),
0.89 (s, 9H, t-Bu), 0.09 (s, 3H, CH3Si), 0.07 (s, 3H, CH3Si); 13C NMR
(75.5MHz, CDCl3) δ 209.2, 167.6, 84.1, 65.6, 63.4, 57.8, 55.0, 50.2,
34.4, 31.4, 25.7, 22.8, 17.8, -4.5, -4.8; MS (FAB, THIO), m/e (rel
intensity) 356.2 (MH+, 14), 298.4 (25), 181.2 (18), 156.2 (100); HRMS
calcd for C18H34O4NSi: 356.2257, found 356.2252.
Benzyl {(3S,4R)-3-[(1R)-1-((tert-butyldimethylsilyl)oxy)ethyl]-4-
[(1S,3S)-3-methoxy-2-oxocyclohexyl]-2-oxoazetidin-1-yloxoacetate
(15). To a solution of 4 (355 mg, 1.0 mmol) in CH2Cl2 (2 mL), was
added a solution of benzyl oxalyl chloride (297 mg, 1.5 mmol) in CH2-
Cl2 (2 mL) and pyridine (160 µl, 2.0 mmol) at 0 °C. The reaction was
warmed to room temperature, stirred for 3 h, poured into water (20
mL), and extracted three times with CH2Cl2 (25 mL). The combined
organic layers were dried over MgSO4, filtered, and concentrated.
Purification by chromatography (hexanes-ethyl acetate, 4:1) afforded
the desired product 15 (500 mg, 97%) as a colorless oil; [R]D -42.3°
(c 1.40, CHCl3); IR (CHCl3) 2940, 2880, 1820 (CO), 1760 (CO), 1740
(CO), 1700 (CO), 1400, 1110 cm-1; MS (FAB), m/e (rel intensity)
518.4 (MH+, 15), 486.4 (43), 460.3 (100), 159.2 (72); HRMS calcd
for C27H40O7NSi 518.2574, found 518.2592.
Preparation of 4 from the Epimerization of a 1:1 Mixture of 4
and 13. Following procedures described above, a ca. 1:1 mixture of
4 and 13 (1.47 g, 69%) was obtained from 7 (1.34 g, 6.3 mmol) and
6 (1.72 g, 6.0 mmol) after chromatography (hexanes-ethyl acetate,
4:1 to 0:1) to remove 14 (235 mg, 11%). The 1:1 mixture of 4 and 13
was subjected to epimerization as follows. To a solution of LDA
prepared by adding n-BuLi (3.4 mL, 8.5 mmol, 2.5 M in hexanes) to
a solution of diisopropylamine (1.3 mL, 9.3 mmol) in THF (10 mL) at
0 °C and stirring for 15 min, was added dropwise a solution of 4 and
13 (ca. 1:1, 1.0 g, 2.8 mmol) in THF (10 mL) over 15 min at -10 °C.
After an additional 10 min, a solution of ZnBr2, prepared by refluxing
a solution of 1,2-dibromoethane (0.26 mL, 3.0 mmol) in THF (10 mL)
over zinc powder (390 mg, 6.0 mmol) for 2 h, was added via cannula,
and the reaction was allowed to warm to 0 °C over 10 min. Diethyl
malonate (2.2 mL, 14 mmol) was added, and the reaction was warmed
to room temperature over 30 min and stirred for an addition 10-15
min (premature workup leads to lower ratios). The reaction mixture
was then poured into saturated aqueous NH4Cl (50 mL) and extracted
twice with ethyl acetate (100 mL). The aqueous phase was made acidic
by the addition of 10% HCl (10 mL) and extracted with ethyl acetate
(100 mL). The combined organic layers were washed with brine (50
mL), dried over MgSO4, and filtered, and the solvents were removed.
Recrystallization of the residue from hexanes afforded 4 as a white
crystalline solid mp 124-126 °C (510 mg, 51%). Chromatography of
the mother liquor (hexane-ethyl acetate, 2:1) afforded an additional
quantity (210 mg, 21%) of 4 as a white crystalline solid (total yield
Benzyl (4S,8S,9R,10S)-4-Methoxy-10-[(1R)-1-((tert-butyldimeth-
ylsilyl)oxy)ethyl]-11-oxo-1-azatricyclo[7.2.0.03,8]undec-2-ene-2-car-
boxylate (16). To a solution of 15 (485 mg, 0.94 mmol) in xylene (5
mL) was added triethyl phosphite (0.8 mL, 4.7 mmol). The resulting
solution was heated to 155-160 °C for 14 h and then cooled to room
temperature. The reaction mixture was then poured onto a silica gel
column and eluted with hexanes to remove the xylene, followed by a
hexanes-ethyl acetate (4:1) eluant to obtain the desired product 16
(378 mg, 83%) as a colorless oil: [R]D 50.5° (c 1.26, CHCl3); IR
(CHCl3) 2940, 2860, 1775 (â lactam CO), 1740 (ester CO), 1670, 1290,
1140 cm-1 1H NMR (300 MHz, CDCl3) δ 7.47-7.28 (m, 5H,
;
aromatic), 5.35 (d, 1H, J ) 12.6 Hz, benzylic H), 5.23 (d, 1H, J )
12.6, benzylic H), 4.96 (apparent t, 1H, MeOCH), 4.27-4.21 (m, 1H,
TBSOCH ), 4.16 (dd, 1H, J ) 3.3, 10.5 Hz, â lactam R to N), 3.50-
3.15 (m, 2H, HC8 and HC9), 3.24 (s, 3H, CH3O), 2.10-2.03 (m, 1H,
cHex), 1.88-1.81 (m, 2H, cHex), 1.69-1.27 (m, 3H, cHex), 1.24 (d,
3H, J ) 6.2 Hz, CH3CHOTBS), 0.88 (s, 9H, tBuSi), 0.08 (s, 6H, Me2-
Si); 13C NMR (75.5MHz, CDCl3) δ 175.6, 160.9, 148.5, 135.4, 128.3,
128.0, 127.8, 126.3, 72.1, 66.6, 65.9, 61.0, 56.0, 54.8, 43.8, 32.3, 30.6,
25.6, 22.3, 20.2, 17.8, -4.3, -5.0 ppm; MS (FAB), m/e (rel intensity)
486.3 (MH+, 40), 428.2 (65), 286.1 (46), 159.1 (100); HRMS calcd
for C27H40O5NSi 486.2676, found 486.2716.
1
72%, H NMR of crude reaction mixture showed an R-â MeO ratio
of 21/1).
Benzyl (4S,8S,9R,10S)-4-Methoxy-10-[(1R)-1-hydroxyethyl]-11-
oxo-1-azatricyclo[7.2.0.03,8]undec-2-ene-2-carboxylate (17). To a
solution of 16 (429 mg, 0.88 mmol) in THF (10 mL) were added acetic
acid (230 µL, 2.7 mmol) and a solution of tetrabutylammonium fluoride
(3 mL, 3 mmol, 300 MHz1 M in THF). The reaction mixture was
stirred for 4 days at room temperature and then diluted with ethyl ether
(300 mL). The organic layer was washed with saturated aqueous
NaHCO3 (50 mL), saturated aqueousNH4Cl (2 × 50 mL), and brine
(50 mL), dried over MgSO4, and filtered. After removal of the solvents,
the crude product was purified by chromatography (hexanes-ethyl
acetate, 1:1) to afford the desired product 15 (288 mg, 88%) as colorless
oil: [R]D +66.8° (c 2.82, CHCl3); IR (CHCl3) 2940, 2880, 1770 (CO,
General Procedure for the Protonation of the Lithium Enolates
of 4 and/or 13. To a solution of LDA (0.38 mL, ca. 0.085 mmol),
prepared by adding n-BuLi (0.4 mL, 1 mmol, 2.5 M in hexanes) to a
solution of diisopropylamine (0.15 mL, 1.1 mmol) in THF (4 mL) at
0 °C and stirring for 15 min, was added dropwise over 1 min a solution
of a mixture of 4 and 13 (ca 1:1, 10 mg, 0.028 mmol) in THF (1 mL).
After stirring for 30 min, a solution of the proton source (5 equiv) in
THF was added all at once. The solution was warmed to room
temperature and then poured into ethyl acetate (50 mL) and saturated
aqueous ammonium chloride (10 mL). The layers were separated, and
the organic layer was washed with brine, dried over Na2SO4, and
concentrated in vacuo. Integration of the methoxy CH3 resonances (â-
methoxy at 3.29 ppm, R-methoxy at 3.46 ppm) in the 1H NMR spectra
of the crude reaction mixtures gave the ratios listed in Table 1.
â-lactam), 1720 (CO, benzyl ester), 1640, 1380, 1090 cm-1 1H NMR
.
(300 MHz, CDCl3) δ 7.46-7.27 (m, 5H, aromatic), 5.37 (d, 1H, J )
12.5 Hz, benzylic), 5.21 (d, 1H, J ) 12.5 Hz, benzylic), 4.94 (apparent
t, 1H, MeOCH), 4.26-4.22 (m, 1H, HOCH), 4.18 (dd, 1H, J ) 3.2,
10.2 Hz, â lactam R to N), 3.28-3.19 (m, 2H, HC8 and HC9), 3.20 (s,
3H, CH3O), 2.07-2.01 (m, 1H, cHex), 1.90-1.67 (m, 2H, cHex), 1.67-
1.61 (m, 1H, cHex), 1.48-1.26 (m, 2H, cHex), 1.32 (s, 3H, J ) 6.3
Hz, H3CCOH) ppm; 13C NMR (75.5 MHz, CDCl3) δ 175.6, 161.0,
149.4, 135.3, 128.5, 128.2, 128.0, 126.1, 72.3, 66.9, 65.5, 60.4, 56.0,
54.9, 44.0, 32.4, 30.6, 21.5, 20.1 ppm; MS (FAB, NBA), m/e (rel
intensity) 372.2 (MH+, 20), 340.4 (20), 286 (20), 167.2 (22), 149.1
(100); HRMS calcd for C21H26O5N 372.18109, found 372.1824.
(4S,8S,9R,10S)-4-Methoxy-10-[(1R)-1-hydroxyethyl]-11-oxo-1-
azatricyclo [7.2.0.03,8]undec-2-ene-2-carboxylic Acid Amidinium Salt
(19). To a solution of 17 (241 mg, 0.65 mmol) in 1,4-dioxane (4 mL)
were added 10% palladium-on-carbon (200 mg) and 3,3,6,9,9-pentam-
ethyl-2,10-diazabicyclo[4.4.0]dec-1-ene (146 µL, 0.65 mmol). The
resulting mixture was stirred under 1 atm of hydrogen for 1 h at room
General Procedure for the Protonation or Deuteration of the
Zinc Enolate of 4 and/or 13. To a solution of the lithium enolate
prepared as above in THF was added at -78 °C a solution of zinc
bromide (0.14 mL, 0.028 mmol, 0.2 M in THF). Stock solutions of
0.2 M zinc bromide were prepared by refluxing 1,2-dibromoethane (172
µL, 2.0 mmol) over zinc powder (320 mg, 5 mmol) for 2 h and diluting
to a total volume of 10 mL with THF. The reaction mixture was
warmed to 0 °C, and the proton source (5 equiv) was added either neat
if it was a liquid or as a solution in THF if a solid. The reaction was
then warmed to room temperature and stirred for approximately 30
1
min. Workup was as described above, and H NMR analysis of the
crude mixtures gave the ratios listed in Tables 2 and 4. Deuteration
experiments were conducted in a similar manner. The additives were
added to the zinc enolate at 0 °C, and the solution was stirred at room
temperature for the specified times and workedup with ethyl acetate

Total Synthesis of 4-Methoxytrinem
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9891
temperature and then filtered through Celite. The solution of the
amidinium salt was the lyophilized to obtain the desired product 19
(318 mg, 100%) as a white amorphous solid: [R]D +60.4° (c 1.03,
CHCl3); IR (CHCl3) 2930, 1770 (â-lactam CO), 1670 (carboxylate CO),
supernatant was decanted. The residue was suspended in diethyl ether
(1 mL), stirred for 1 min, and centrifuged, and the supernatant was
decanted. This process was repeated three times, and the product was
dried in vacuo to afford the title compound (24 mg, 54%) as a white
amorphous solid. A 1H NMR spectrum (300 MHz, D2O) was identical
to that of an authentic sample provided by Glaxo-Wellcome (Verona,
Italy).
1400, 1150 cm-1 1H NMR (300 MHz, CDCl3) δ 11.35 (br, 2H, 2
.
amidinium NH), 5.27 (t, 1H, J ) 2.8 Hz, MeOCH), 4.24-4.19 (m,
1H, HOCH), 4.09 (dd, 1H, J ) 3.1, 10.2 Hz, â-lactam R to N), 3.28
(s, 3H, CH3O), 3.16 (dd, 1H, J ) 3.1, 6.9 Hz, â-lactam R to CO),
3.12-3.04 (m, 1H, C8), 2.04-1.90 (m, 4H), 1.89 (m, 6H), 1.59-1.50
(m, 4H), 1.39 (s, 6H), 1.32-1.30 (m, 12H) ppm; 13C NMR (75.5MHz,
CDCl3) δ 175.5, 167.8, 166.8, 139.1, 133.4, 72.6, 66.9, 65.9, 59.8, 55.6,
55.0, 53.3, 53.2, 43.3, 32.3, 31.5, 31.3, 31.2, 30.8, 30.7, 30.3, 29.8,
29.7, 24.3, 21.8, 20.6 ppm.
Acknowledgment. We thank the National Science and
Engineering Council of Canada for generous financial support.
This research was part of a technology transfer program with
Glaxo-Wellcome (Verona, Italy).
Supporting Information Available: 1H and 13C NMR
spectral data for most of the compounds in this paper (23 pages).
See any current masthead for ordering and Internet access
instructions. Ordering information is given on any current
masthead page.
(4S,8S,9R,10S)-4-Methoxy-10-[(1R)-1-hydroxyethyl]-11-oxo-1-
azatricyclo[7.2.0.03,8]undec-2-ene-2-carboxylic Acid Potassium Salt.
To a solution of the amidinium salt 19 (71 mg, 0.15 mmol) in ethyl
ether (1 mL) and ethyl acetate (1 mL) was added at room temperature
a solution of potassium 2-ethylhexanoate (310 µL, 0.16 mmol, 0.51 M
in ethyl acetate). The resulting suspension of potassium salt was stirred
for 15 min and placed in a centrifuge for 5 min, after which the
JA962006N