Stereoselective synthesis of a key intermediate of sanfetrinem by means of a chelated tin(IV) enolate

Article abstract of DOI:10.1021/jo962095m

(2S)-2-Methoxycyclohexanone [(2S)-6], reacts with the 4-acetoxyazetidinone 3a in the presence of SnCl₄ and a tertiary amine base (such as N,N-diisopropylethylamine) to give the ketoazetidinone 5 (a key intermediate in the synthesis of the broad

Full text of DOI:10.1021/jo962095m

J . Org. Chem. 1997, 62, 1653-1661
1653
Ster eoselective Syn th esis of a Key In ter m ed ia te of Sa n fetr in em by
Mea n s of a Ch ela ted Tin (IV) En ola te
Tino Rossi,* Carla Marchioro, Alfredo Paio, Russell J . Thomas, and Paola Zarantonello
Glaxo Wellcome S.p.A. Medicines Research Centre, Via Fleming 4, 37100 Verona, Italy
Received November 8, 1996X
(2S)-2-Methoxycyclohexanone [(2S)-6], reacts with the 4-acetoxyazetidinone 3a in the presence of
SnCl4 and a tertiary amine base (such as N,N-diisopropylethylamine) to give the ketoazetidinone
5 (a key intermediate in the synthesis of the broad-spectrum antibiotic sanfetrinem 2a ) with high
yield and diastereoselectivity. Low-temperature NMR studies of the reaction indicated the formation
of a 1:1 chelate complex 11 between SnCl4 and (2S)-6 which, on addition of the base is transformed
into the highly reactive chelated tin enolate 12. The formation of compound 11 has been confirmed
by single-crystal X-ray analysis. The high diastereoselectivity observed is believed to derive from
an open transition state where the chelated SnCl4 amplifies the stereochemical influence of the
methoxy group. This reaction offers considerable advantages over all existing syntheses of the
ketoazetidinone 5 and is currently under evaluation for inclusion in the industrial synthesis of
sanfetrinem.
In tr od u ction
In recent years our interest on the synthesis of new
antibacterial agents has been concentrated on a new class
of antibiotics, the trinems (1), first disclosed in our
laboratories,1 which demonstrated extremely interesting
biological properties both in vitro and in vivo. San-
fetrinem (GV104326, 2a )2 (Figure 1) and its metabolically
labile ester 2b have been selected for development studies
and are currently in phase II clinical trials.
Due to the complexity of the structure, bearing five
F igu r e 1.
stereogenic centers, and the large amounts of final drug
material required to support development studies, a high-
yielding, short, practical, and robust route to 2a and 2b
was required. Although the existing route1c proved to
be robust and provided us with multikilogram quantities
of the desired trinems to support early development
studies, it suffered from two major drawbacks: (a) low
overall chemical yields (8-9%); (b) high number of stages
(11 steps, 15 chemical transformations).
Sch em e 1
We have already reported our results in the direct
synthesis of two key intermediates 43,4 and 55,6 from
commercially available acetoxyazetidinone 3a (Scheme
1). In both approaches significant advances have been
made both in terms of yields and in the reduction of
number of stages. Although a convergent and stereose-
lective synthesis of ketoazetidinone 5 could be conve-
niently established, the procedure was still rather com-
X Abstract published in Advance ACS Abstracts, February 1, 1997.
(1) (a) Tamburini, B.; Perboni, A.; Rossi, T.; Donati, D.; Andreotti,
D.; Gaviraghi, G.; Carlesso, R; Bismara C. Eur. Pat. Appl. EP0416953
A2, 1991; Chem. Abstr. 1992, 116, 235337t. (b) Perboni, A.; Rossi, T.;
Gaviraghi, G.; Ursini, A.; Tarzia G. WO 9203437, (c) Perboni, A.;
Tamburini, B.; Rossi, T.; Donati, D.; Tarzia G.; Gaviraghi, G. In Recent
Advances in the Chemistry of Anti-Infective Agents; Bentley, P. H.,
Ponsford, R., Eds.; RSC: Cambridge 1992; pp 21-35. (d) Andreotti,
D.; Rossi, T.; Gaviraghi, G.; Donati, D., Marchioro, C.; Di Modugno,
E.; Perboni, A. D. Bioorg. Med. Chem. Lett. 1996, 6, 491.
(2) Di Modugno, E.; Erbetti, I.; Ferrari, L.; Galassi, G. L.; Hammond,
S. M.; Xerri, L. Antimicrob. Agents Chemother. 1994, 38, 2362.
(3) Bismara, C.; Di Fabio, R.; Donati, D.; Rossi, T.; Thomas, R. J .
Tetrahedron Lett. 1995, 36, 4283.
plicated by the fact that in order to ensure high yield
and good diastereoselectivity, a protection-deprotection
step of the azetidinone nitrogen was also required.5
Moreover, enantiomerically pure silyl enol ether deriva-
tive of (2S)-6, required for this coupling, could be obtained
with high regioselectivity only under strictly controlled
conditions, employing the expensive trimethylsilyl trif-
luoromethanesulfonate.
An alternative linear route to 5 has also recently been
published7 but was not considered amenable to an
industrial synthesis of sanfetrinem.
A less complicated procedure for the synthesis of 5 from
3a was highly desirable, and we decided to turn our
(4) Rossi, T.; Biondi, S.; Contini, S.; Thomas, R. J .; Marchioro, C. J .
Am. Chem. Soc. 1995, 117, 9604.
(5) Ghiron, C.; Piga, E.; Rossi, T.; Tamburini B.; Thomas, R. J .
Tetrahedron Lett. 1996, 37, 3891.
(6) Kennedy, G.; Rossi, T.; Tamburini, B. Tetrahedron Lett. 1996,
37, 7441.
(7) Hanessian, S.; Rozema M. J . J . Am. Chem. Soc. 1996, 118, 9884.
S0022-3263(96)02095-6 CCC: $14.00 © 1997 American Chemical Society

1654 J . Org. Chem., Vol. 62, No. 6, 1997
Rossi et al.
Ta ble 1. In flu en ce of th e Rea ction Tem p er a tu r e a n d
Effect of Differ en t Lew is Acid sa
Sch em e 2
Lewis acid
(equiv)
5 yieldb ratio
(%)
entry
base (equiv)
T (°C)
5:7c
1
2
3
4
SnCl4 (3)
SnCl4 (3)
Et3N (2.2)
Et3N (2.2)
Et3N (2.6)
Et3N (2.6)
Et3N (2.6)
Et3N (2.6)
Et3N (2.6)
-78 to 0
-20 to 0
-20 to 0
-20 to 0
-20 to 0
0 to 10
35
39
50
38
0
13:1
16:1
10:1
10:1
SnCl4 (3.4)
SnBr4 (3.4)
TiCl4 (3.4)
SnCl4 (3.4)
SnCl4 (3.4)
5
6
60
44
8.3:1
30:1
7d
0 to 10
a
Reaction in anhydrous CH2Cl2; 1 g of 3 and 2 equiv of (2S)-6
were used; (2S)-6 was purified by flash chromatography prior to
b
use. Solution yield determined by HPLC (Hypersil ODS2, 250
mm × 4 mm × 5 µm, buffer (NH4)H2PO4 50 mM/CH3CN 45/55,
flow rate 1.0 mL/min, UV at 205 nm, 5 as external standard).
d
c Ratios determined by NMR. Compound 3b was employed.
give any detectable reaction product. Increasing reaction
temperature and variation of reactant ratios did not
result in any improvement and, at temperatures above
-10 °C extensive decomposition of 3a was observed. This
was in line with previous observations made during our
studies in the Lewis acid-mediated coupling of silyl enol
ethers with 3a and its N-trimethylsilyl derivative 3b.5
The use of TiCl4 always led to poor results and extensive
decomposition of the 4-acetoxyazetidinone.
Sch em e 3
On the basis of these results we decided to replace
TiCl4 with another Lewis acid that could efficiently
activate 3a under the reaction conditions. Tin(IV) chlo-
ride was considered a possible candidate because it had
already demonstrated its efficacy in the one-pot formation
of 5.5
The availability of sufficient amounts of enantiomeri-
cally pure (2S)-65,11 led us to initiate our studies with this
reactant in order to reduce the complexity of the reaction
mixture by reducing the number of the possible isomers.
Initial experiments were made under standard protocol
conditions (generation of the enolate at -78 °C followed
by addition of either 3a or a preformed mixture of 3a
and a Lewis acid), but we were able to detect only traces
of coupling products 5 and 7 (HPLC and NMR analysis
of the crude reaction mixtures). However, variation of
the reaction conditions led us to find a simpler and more
satisfactory protocol: addition of triethylamine to a
mixture of 3a , (2S)-6, and SnCl4, respectively at -78 °C
and then raising the temperature to 0 °C resulted in the
formation of a mixture of products in which the presence
of compounds 5 and 7 could be measured (HPLC and
NMR). The results of these initial experiments are
reported in Table 1. We were pleased to observe that
the ratio of the two isomers originated by the same
enantiomer (2S)-6 (5 and 7) was high and in favor of the
desired ketoazetidinone 5. Increasing the reaction tem-
peratures resulted in a significant increase in solution
yield although a reduction in stereoselectivity was ob-
served. Under these new conditions TiCl4 again failed
to give the desired reaction products while SnBr4 gave
similar results to SnCl4. The use of N-silylated derivative
3b gave a reduction in yields and, as expected, an
improvement in terms of stereoselectivity (compare en-
tries 6 and 7). Due to the more complicated conditions
that have to be employed when working with 3a we
decided to carry out all subsequent studies on 3b . A
attention to the direct coupling between 2-methoxycy-
clohexanone (6) and acetoxyazetidinone 3a .
Tin En ola te Ch em istr y
It is well known that titanium(IV) chloride can promote
base-mediated enolization of ketones.8 Treatment of
titanium enolates thus formed with an aldehyde leads
to the formation of aldol condensation products. This
protocol has been applied by Cozzi and co-workers in the
stereoselective synthesis of â-lactams from chelated
thiopyridyl ester enolates and imines.9
A 1:1 complex between 2-methoxycyclohexanone and
TiCl4 has been the subject of NMR studies by Reetz and
co-workers.10 We reasoned that an enolate formed from
this chelate complex could give preferentially the re-
quired diastereomer with good diastereoselectivity in the
condensation reaction with 3a .
We decided to initiate our studies on the direct coupling
between the titanium enolate of 2-methoxycyclohexanone
(6) and 4-acetoxyazetidinone 3a (Scheme 2). Thus,
treatment of 2 equiv of a 1:1 mixture of 6 and TiCl4 with
triethylamine in anhydrous dichloromethane at -78 °C
followed by addition of 1 equiv of 3a at -78 °C did not
(8) Evans D. A.; Rieger, D. L.; Bilodeau, M. T.; Urpi, F. J . Am. Chem.
Soc. 1991, 113, 2047.
(9) (a) Annunziata, R.; Benaglia, M.; Cinquini, M.; Cozzi, F.; Ponzini,
F. J . Org. Chem. 1993, 58, 4746. (b) Cinquini, M.;Cozzi, F.; Cozzi, P.
G.; Consolandi, E. Tetrahedron 1991, 47, 8767.
(10) Reetz, M. T.; Kesseler, K.; Schimdtberger, S.; Wenderoth, B.;
Steinbach, R. Angew. Chem., Int. Ed. Engl. 1983, 22, 989.
(11) Stead, P.; Marley H.; Mahmoudian, M.; Webb, G.; Noble, D.;
Ip, Y. T.; Piga, E.; Rossi, T.; Roberts, S.; Dawson, M. J . Tetrahedron
Asymmetry 1996, 7 (8), 2347.

Synthesis of a Key Intermediate of Sanfetrinem
J . Org. Chem., Vol. 62, No. 6, 1997 1655
F igu r e 2. 1H-NMR at -40 °C: (a) 6; (b) 6 + 1 equiv of SnCl4; (c) 6 + 1 equiv of SnCl4 + 1 equiv of lutidine.
Ta ble 2. In flu en ce of Ba ses a n d Solven tsa
peratures of 0-5 °C gave consistently the best results
while it was found that reduction of the equivalents of
(2S)-6 did not result in a dramatic decrease in yield and
gave a substantial reduction in side products of the
reaction (Table 2, entry 7).
ratio
entry solvent
base (equiv)
T (°C)
5 (yield)b
5:7c
1
2
PhCl
PhF
Et3N (2.6)
Et3N (2.6)
-20 to 0
-20 to 0
-20 to 0
-5 to 0
49%
54%
5%
65%
77%
60%
70%
22:1
11:1
nd
15:1
11:1
25:1
18:1
3
4
5
PhCH3 Et3N (2.6)
CH2Cl2 i-Pr2EtN (2.6)
CH2Cl2 i-Bu3N (2.6)
NMR Stu d ies
- 5 to 0
6
CH2Cl2 2,6-lutidine (2.6) -5 to 0
CH2Cl2 i-Bu3N (2.6) -10 to 0
In order to gain an insight into the reaction mecha-
nism, variable-temperature NMR studies were under-
taken.
7d
a
One gram of 3, 2 equiv of (2S)-6 and 3.4 equiv of SnCl4 were
used; (2S)-6 was purified by flash chromatography prior to use.
A first set of experiments on 2-methoxycyclohexanone
(6) were carried out at various temperatures between
-40 and 23 °C in CD2Cl2.12 The addition of measured
amounts of SnCl4 to a solution of 6 in CD2Cl2 resulted in
a dramatic change of the 1H-,13C-, and 119Sn-NMR
spectra. In particular, after the addition of 1 molar equiv
of tin(IV) chloride, the complete transformation of 6 into
a new species was observed (Figures 2-4). In the proton
spectrum, at -40 °C, all the proton signals were shifted
downfield to varying degrees; in particular, the hydrogens
at C-2, C-6, and those of the methoxy group showed the
most significant changes in chemical shift: the proton
at C-2 moves from 3.80 to 4.58 ppm with a change also
in the coupling constant values, suggesting modification
of the ring conformation, similarly, the chemical shift of
b
Solution yield determined by HPLC (5 as external standard).
c Ratios determined by NMR. A total of 1.2 eq of (2S)-6 were used.
d
number of experiments with racemic 6 were also carried
out; mixtures of diastereomers 5 and 7-9 were always
observed. Ketoazetidinones 5 and 8 (originated by the
two enantiomers of 6) were the major isomers and their
ratio by NMR of the crude reaction mixture was always
close to 1:1, indicating no difference in reactivity between
the two enantiomers.
Encouraged by these promising initial observations,
studies on the reaction conditions were immediately
undertaken. The influence of the base, reagent propor-
tions, temperature, and solvents were analyzed and some
representative results are reported in Table 2. Among
the amines studied, triethylamine, N,N-diisopropylethy-
lamine, triisobutylamine and 2,6-lutidine were found to
be effective and high solution yields were measured.
Dichloromethane and chlorobenzene, with reaction tem-
(12) Conformational studies on 2-methoxycyclohexanone have al-
ready been published, see: (a) Fraser, R. R.; Faibish, N. C. Can. J .
Chem. 1995, 73, 88. (b) Basso, E.; Kaiser, C.; Rittner, R.; Lambert, J .
B. J . Org. Chem. 1993, 58, 7865.

1656 J . Org. Chem., Vol. 62, No. 6, 1997
Rossi et al.
F igu r e 3. 13C-NMR at -40 °C: (a) 6; (b) 6 + 1 equiv of SnCl4; (c) 6 + 1 equiv of SnCl4 + 1 equiv of lutidine.
the methoxy group moves from 3.44 to 3.97 ppm. In the
13C-NMR, carbons C-1 and C-2 and the methoxy group
again showed the most significant change in chemical
shifts (Figure 3): C-2 moves from 86.0 to 89.0 ppm, the
carbonyl carbon undergoes a notable shift downfield from
212.1 to 240.3 ppm, and the methoxy carbon moves from
59.7 to 57.5 ppm. In the 119Sn spectrum, a single signal
at -542 ppm was observed, suggesting the formation of
a hexacoordinate tin species.13,14 On the basis of this
evidence the formation of a 1:1 chelate complex (11,
Figure 2) between 6 and SnCl4 was proposed. This
complex was found to be stable in CD2Cl2 at room
temperature for at least one week.
In a third series of experiments a 1:1 solution of 6 and
SnCl4 in CD2Cl2 was cooled to -40 °C and treated with
1 equiv of 2,6-lutidine. This base was selected both
because it proved to work efficiently in the SnCl4-
promoted reaction between 3a and 6 and for its low
interference in the analysis of NMR spectra.
1
On recording the H-,13C-, and 119Sn-NMR spectra at
-40 °C showed the formation of a new species (Figures
2-4). Protons at C-2 and of the methoxy group were
found to undergo a moderate shift upfield compared to
11: the proton at C-2 moves from 4.58 to 4.35 ppm,
protons of the methoxy group move from 3.97 to 3.85 ppm
while disappearance of signals belonging to hydrogens
at C-6 and appearance of a new signal at 4.69 ppm
indicated the formation of enolate 12. 13C- and 119Sn-
NMR spectra were also in agreement with the proposed
structure with C-6 giving a signal at 112.9 ppm. The
119Sn-NMR spectrum showed a single peak at -546 ppm,
again indicating the presence of a hexacoordinate tin
atom. Intermediate 12 was stable at low temperatures
(-40 °C) but degradation was found to occur rapidly at
0 °C.
Further support to our proposed structure (11) came
from the literature. We have already mentioned that a
1:1 chelate complex between 6 and TiCl4 has been
reported;9 moreover formation of stable 1:1 chelate
complex between SnCl4 and R-alkoxyaldehydes15 and
ketones11 have been published.
In separate experiments 0.5 and 2 equiv respectively
of SnCl4 were added to 6 . In the first case a mixture of
starting material 6 and complex 11 was measured while
in the second case an excess of tin(IV) chloride did not
result in detectable changes to the NMR spectra.
A similar series of experiments were also run on
1
4-acetoxyazetidinone 3a . The H- and 13C-NMR spectra
of 3a were first recorded at -40, 0, and 23 °C. Addition
of measured amounts of SnCl4 led to a significant change
(13) (a) Reetz, M. T.; Harms, K.; Reif, W. Tetrahedron Lett. 1987,
28, 4515. (b) Reetz, M. T. Acc. Chem. Res. 1993, 26, 462.
(14) Denmark S. E.; Almstead, N. G. J . Am. Chem. Soc. 1993, 115,
3133.
1
in the H- and 13C-NMR spectra (as shown in Figures 5
and 6). In particular, after addition of 0.5 molar equiv
of SnCl4 the disappearance of signals corresponding to
(15) Keck, G. E.; Castellino, S. J . Am. Chem. Soc. 1986, 108, 3847.

Synthesis of a Key Intermediate of Sanfetrinem
J . Org. Chem., Vol. 62, No. 6, 1997 1657
F igu r e 4. 119Sn-NMR at -40 °C: (a) 6 + 1 equiv of SnCl4; (b) 6 + 1 equiv of SnCl4 + 0.5 equiv of lutidine; (c) 6 + 1 equiv of SnCl4
+ 1 equiv of lutidine.
3a and the formation of a new species was observed.
Further addition of SnCl4 did not modify the NMR
spectra. The 119Sn-NMR spectra, recorded at -60 °C
(Figure 7), showed a sharp peak at -645 ppm. On the
basis of these initial observations one only could hypoth-
esize, in agreement with previous observation by Den-
mark on NMR studies on benzaldehyde-SnCl4 com-
plexes,12 the formation of a 2:1 complex between 3a and
SnCl4. However a clear idea of the structure of the
complex could not be inferred. Analysis of the 13C-NMR
spectra with heteronuclear correlation experiments per-
mitted the precise assignment of CdO signals, leading
us to establish that 3a is complexed with SnCl4 via the
lactam carbonyl oxygen atom (13, Figure 5). This experi-
ment allowed us to assign the signal at 168.7 ppm to the
lactam carbon which, upon addition of SnCl4, undergoes
a dramatic shift from 168.4 to 177.6 ppm. This complex
is rather unstable at temperatures above 0 °C but
nevertheless satisfactory NMR spectra could also be
recorded for a short time at room temperature.
NMR reaction mixture, after quenching with bicarbonate
and Rochelle’s salt, showed the formation of the desired
product 5 together with a small amount of its isomer 7
and starting materials 3a and (2S)-6.
Isola tion a n d Str u ctu r e Elu cid a tion of th e
Tin (IV) Ch lor id e Ch ela te Com p lex 11
During our studies on this reaction we noticed that
addition of SnCl4 to a mixture of 3a and 6 resulted in
the formation of a white precipitate that upon addition
of base, redissolved, giving a homogeneous solution.
Addition of SnCl4 to a solution of (2S)-6 in chloroben-
zene resulted in a precipitation of an off-white solid that
could be isolated by filtration under inert atmosphere and
could be stored in a closed vessel at room temperature
for several weeks. NMR spectra recorded for this com-
pound were identical to those assigned to compound 11.
Crystals of this compound were obtained by recrystalli-
zation from chlorobenzene under an inert atmosphere
and X-ray analysis confirmed the proposed structure 11
(Figure 8). This isolated complex was successfully used
in the synthesis of 5, giving similar results in terms of
both yield and stereoselectivty to those obtained by
employing 3a and (2S)-6, under similar reaction condi-
tions. In Table 3 a direct comparison between experi-
ments with (2S)-6 and 11 is made.
In a final experiment 2 equiv of enolate 12 (generated
from (2S)-6 as described above in CD2Cl2 at -70 °C) were
added to the preformed complex 13 at -40 °C. A proton
NMR at -40 °C was recorded and showed an unreacted
mixture of the two components. The reaction mixture
was then warmed rapidly to 0 °C: NMR analysis showed
the disappearance of 12 and the formation of a complex
mixture of new products. HPLC analysis of the crude

1658 J . Org. Chem., Vol. 62, No. 6, 1997
Rossi et al.
F igu r e 5. 1H-NMR at -40 °C: (a) 3; (b) 3 + 0.5 equiv of SnCl4.
Ta ble 3. Use of Ch ela te Com p lex 10
for the facial selectivity on the azetidinone ring and the
bulky chelated tin tetrachloride amplifying the steric
effect of the methoxy group on 6 and preferentially
orientating the elecrophile to attack 12 from the less
hindered opposite face.
It is worth noting that under our best reaction condi-
tions, the presence of regiosomers 14 and 15 (Figure 10)1d,
could not be detected (NMR and HPLC) in the crude
reaction mixture, thus demonstrating that the regiose-
lective enolization of 6, as observed during low-temper-
ature NMR experiments, does occur even at 0 °C.
Moreover, the absence of diastereomers 8 and 9 in the
crude reaction mixture, when enantiomerically pure
(2S)-6 was used, indicates that reaction conditions are
mild enough to prevent racemization at the stereogenic
center of 6.
ratio
entry method solvent base (equiv) T (°C) 5 yieldb 5:7c
1
2
3
4
A
A
B
B
CH2Cl2 i-Bu3N (2.6) -10 to 0
70%
64%
79%
66%
18:1
17:1
23:1
15:1
PhCl
PhCl
i-Bu3N (2.6) -10 to 0
i-Bu3N (2.6) -20 to 0
CH2Cl2 i-Bu3N (2.6) -20 to 0
a
Method A. Addition of base to a precooled mixture containing
1 g of 3a , 1.2 equiv of (2S)-6, and 3.4 equiv of SnCl4. Method B.
Addition of base to a precooled mixture containing 1 g of 3a , 1.2
equiv of 10 and 2.2 equiv of SnCl4. (2S)-6 was purified by flash
chromatography prior to use b Solution yield determined by HPLC
(5 as external standard). c Ratios determined by NMR.
Th e Rea ction Mech a n ism
On the basis of NMR experiments an hypothesis on
the reaction mechanism could be drawn. We believe that
enolate 12 is the actual reacting species responsible for
the nucleophilic attack on azetinone A (Figure 9), gener-
ated from 3a under the influence of a Lewis acid.16
Although NMR studies have demonstrated that 3a is
coordinated to SnCl4 via its lactam oxygen, the existence,
under the reaction conditions, of other different transient
species that could promote the elimination of the acetoxy
moiety thus generating A cannot be ruled out. The C-C
bond formation occurs via an open transition state
(intermediates I-IV); the stereoselectivity observed is
caused by a double diastereoselective effect with the
silyloxy side chain at C-3 of azetinone A being responsible
P r elim in a r y Op tim iza tion Stu d ies
Further optimization studies led us to conclude that a
reduction in the ratio (2S)-6:3a from 2:1 to 1.2:1 did not
dramatically affect yields but reduced the amounts of
byproducts leading to a simpler purification procedure.
On the other hand all the initial attempts to reduce the
amounts of SnCl4 and tertiary amines were not successful
and, in our hands, the best results were obtained when
3.4 equiv of SnCl4 and 2.6 equiv of base per equivalent
of 3a were employed. N,N-Diisopropylethylamine and
triisobutylamine were the bases of choice and dichlo-
romethane and chlorobenzene gave comparable results.
When the reaction was carried out on a 5-g scale, solution
yields were consistently close to 70% with isolated yield
(16) (a) Endo, M. Can. J . Chem. 1987, 65, 2140. (b) Martel A.; Daris,
J . P.; Bachand, C.; Corbeil, J .; Menard, M. Can. J . Chem. 1988, 66,
1537.

Synthesis of a Key Intermediate of Sanfetrinem
J . Org. Chem., Vol. 62, No. 6, 1997 1659
F igu r e 6. 13C-NMR at -40 °C: (a) 3; (b) 3 + 0.5 equiv of SnCl4.
after crystallization between 60% and 66% (see Experi-
mental Section).
ment toward the synthesis of GV104326. Studies aimed
at a further optimization of the reaction conditions are
currently ongoing and will be reported in due course.
Con clu sion s
Exp er im en ta l Section
We have demonstrated that a simple stereoselective
synthesis of 5 in up to 66% isolated yield (74% solution
yield) could be obtained from (2S)-6 and 3a under mild
conditions by simple addition of a tertiary amine to a
mixture of the two starting materials and SnCl4 in
dichloromethane or chlorobenzene. A high diastereose-
lectivity was observed even without using protecting
groups at the azetidinone nitrogen. The yields obtained
are higher compared to the previously reported method,5
and a much simpler and economic procedure was estab-
lished. This method appears to be a significant improve-
All reactions were performed in oven-dried glassware under
a positive pressure of nitrogen. (1S,2S)-2-Methoxycyclohex-
anol was synthesized according to the reported procedure.10
Solvents were distilled under a nitrogen atmosphere, P2O5
(CH2Cl2). Triethylamine, N,N-diisopropylethylamine, 2,6-
lutidine, and triisobutylamine were distilled from CaH2.
Analytical thin-layer chromatography (TLC) was performed
on E. Merck silica gel 60 F254 plates (0.25 mm). Compounds
were visualized by dipping in a phosphomolybdic acid solution
followed by heating. Flash chromatography was performed
on E. Merck silica gel (230-400 mesh). Melting points are
uncorrected and were determined with a capillary melting

1660 J . Org. Chem., Vol. 62, No. 6, 1997
Rossi et al.
F igu r e 7. 119S-NMR at -40 °C: (a) 3 + 0.5 equiv of SnCl4; (b) 3 + 1 equiv of SnCl4.
F igu r e 8.
point apparatus. IR spectra were recorded in CDCl3 solution
unless otherwise stated and are reported in wavenumbers
(cm-1). 1H-NMR spectra were recorded at 400 or 500 MHz,
13C-NMR were recorded at 100.57 or at 75.43 MHz: all the
spectra were collected in CDCl3 or CD2Cl2 at 25 °C. In the
1H-NMR spectra, chemical shifts are reported in ppm with
respect to residual CHCl3 at 7.26 or to residual CH2Cl2 at 5.32
downfield from the TMS line while, in the carbon NMR
spectra, the center line (δ 77.0 or 53.4) 13C resonance of CDCl3
or CD2Cl2 was used as internal reference. NMR assignments
are assisted by NOE and 2D techniques. Low-temperature
NMR studies were run at 400 MHz (1H-NMR), 100.6 MHz (13C-
NMR), and 112 MHz (119Sn-NMR) using CD2Cl2 as solvent (in
the latter case tetramethyltin was used as external standard);
the samples were placed in the probe at the appropriate
temperature and allowed to equilibrate for a few minutes prior
to acquisition of a spectrum. HPLC analyses were performed
on a HPLC Hypersil ODS2 column (250 × 4 mm) × 5 µm, using
F igu r e 9. Schematization of the proposed transition states.
a buffer of (NH4)H2PO4 50 mM/CH3CN 45/55 and flow rate of
1.0 mL/min, while monitoring at UV at 205 nm. A pure
sample of 5 was used as external standard. Chiral GC were
performed on a Chiraldex G-TA (γ-cyclodextrin, trifluoro-
acetyl), using FID as a detector. Mass spectra were recorded
in FAB+ mode. All optical rotations [R] values were obtained
in CHCl3 or CH2Cl2 solutions at the sodium D line at 22 °C.
(2S)-2-Meth oxycycloh exa n on e ((2S)-6). Concentrated
sulfuric acid (6.1 mL) was added to a stirred solution of
chromium trioxide (7 g, 70 mmol) in water (50 mL) at 0 °C.
An aliquot of the resulting solution (30 mL) was added
dropwise in 15 min to a solution of (1S,2S)-2-methoxycyclo-
hexanol (1.95 g, 15 mmol) in dichloromethane (15 mL) at 0 °C
under vigorous stirring. The reaction mixture was stirred for
1 h at 0 °C and then quenched by addition of isopropyl acohol
(2.5 mL). The mixture was extracted with dichloromethane
(3 × 50 mL), the combined extracts were washed with a
saturated solution of NaHCO3 (30 mL) and then with brine
(50 mL) and dried over magnesium sulfate. The resulting
mixture was filtered over a pad of celite and evaporated under
reduced pressure at room temperature to give a pale yellow
liquid (1.1g, w/w assay 98%, yield 56%). This material could
be used without any further purification. Purification by flash
chromatography gave a pure sample of the title compound as
a colorless oil: [R]22 ) -112.4° (c ) 2.08, CH2Cl2); ee >99%
D

Synthesis of a Key Intermediate of Sanfetrinem
J . Org. Chem., Vol. 62, No. 6, 1997 1661
extracts were washed in turn with a 10% aqueous solution of
citric acid (150 mL), a saturated solution of NaHCO3 (150 mL),
and brine (150 mL). The organic solution was dried over
magnesium sulfate and concentrated under reduced pressure
to give the crude reaction mixture as an off-white solid (5.59
g, w/w assay by HPLC 68.53%; corrected yield 73.8%).
Crystallization from n-hexane (160 mL) gave 4.1 g of the
title compound (w/w assay by HPLC 98.5%; corrected yield
66.7%). Evaporation of the mother liquors gave 2.15 g of a
semisolid residue (w/w assay by HPLC 19.85%; corrected yield
6.9%).
F igu r e 10.
(GC); 1H-NMR (CDCl3, 400 MHz) δ 3.74 (m, 1H), 3.41 (s, 3H),
2.58-2.46 (m, 1H), 2.34-2.19 (m, 2H), 2.01-1.88 (m, 2H),
1.80-1.58 (m, 3H).
NMR analysis of the crude reaction mixture revealed the
following: molar ratio 5:7 ) 20:1; ratio 5:3a ) 23:1; ratio 5:6
>100:1.
Meth od A. (3S,4R)-3-[(R)-1-[(ter t-Bu tyld im eth ylsilyl)-
oxy]eth yl]-4- [(R)-6′-((S)-2′-m eth oxy)-1′-oxocycloh exyl]-
a zetid in -2-on e (5). To a stirred solution (2.65 g, 21 mmol)
of (2S)-6 in anhydrous dichloromethane (23 mL) at -20 °C
was added SnCl4 (6.95 mL, 59.16 mmol) dropwise. To the
resulting suspension was added a solution of 3a (5 g, 17.4
mmol) in dichloromethane (10 mL), and the resulting mixture
was warmed to 0 °C. A solution of N,N-diisopropylethylamine
(7.9 mL, 45.2 mmol) in dichloromethane (10 mL) was added
over 20 min, maintaining the temperature between 0 and 5
°C. The reaction mixture was stirred for a further 40 min and
then poured onto a 1:1 v/v mixture of saturated sodium
hydrogen carbonate and saturated Rochelle’s salt (300 mL).
Ethyl acetate (150 mL) was added and the mixture was stirred
for 1.5 h. The organic layer was separated, washed with brine,
dried over magnesium sulfate, filtered, and evaporated under
reduced pressure to give an off-white solid that was crystal-
lized from n-hexane (160 mL) to give the title compound (3.8
g, w/w assay by HPLC: 97.5%; corrected yield 60%, mp 116-
117.5 °C). Evaporation of the mother liquors gave an oily
residue (2.45 g; w/w assay by HPLC 21.5%; yield 8.5%). NMR
analysis of the crude reaction mixture revealed the following:
(2S)-2-Meth oxycycloh exa n on e Tin (IV) Ch lor id e 1:1
Ch ela te Com p lex (10). To a stirred solution of (2S)-6 (2.0
g, 15.6 mmol) in dry chlorobenzene (15 mL) cooled at -5 °C
was added SnCl4 (4.4 g, 16.9 mmol) dropwise, at such a rate
to maintain the temperature below 0 °C. The resulting
suspension was stirred for further 15 min prior to filtering at
the pump under an inert atmosphere. The filter cake was
washed with n-hexane (30 mL) and dried in vacuo to give
compound 10 as a white-pale pink solid (5.9 g, 97%): mp 162-
163 °C dec; [R]20 ) -15.1° (c ) 0.935, CD2Cl2); 1H-NMR
D
(CDCl3, 500 MHz) δ 4.12 (m, 1H), 3.74 (s, 3H), 2.78 (m, 1H),
1
2.50 (m, 2H), 2.16 (m, 1H), 2.1-1.7 (m, 4H); H-NMR (CD2-
Cl2, 500 MHz) δ 4.61 (m, 1H), 4.06 (s, 3H), 3.04 (m, 1H), 2.84
(m, 1H), 2.78 (m, 1H), 2.38 (m, 1H), 2.2-2.08 (m, 2H), 1.88
(m, 2H); 13C-NMR (CD2Cl2, 75.4 MHz) δ 220.0, 83.28, 59, 31,
40.05, 33.11, 27.86, 22.90; IR (nujol mull) νmax 1649 cm-1
.
Meth od B. (3S,4R)-3-[(R)-1-[(ter t-Bu tyld im eth ylsilyl)-
oxy]eth yl]-4- [(R)-6′-((S)-2′-m eth oxy)-1′-oxocycloh exyl]-
a zetid in -2-on e (5). To a mixture of 4-acetoxyazetidinone 3a
(1.0 g, 3.47 mmol) and compound 10 (1.6 g, 4.11 mmol) was
added anhydrous dichloromethane (25 mL). The resulting
suspension was cooled to -20 °C with stirring, and SnCl4 (1.98
g, 7.63 mmol) was added over 2 min. The reaction mixture
was warmed to 0 °C, and a solution of triisobutylamine (2.2
mL, 9.02 mmol) in anhydrous dichloromethane (5 mL) was
added via cannula over 20 min. The resulting solution was
stirred for a further 40 min at 0 °C before pouring into a
vigorously stirred mixture of saturated Rochelle’s salt solution
(75 mL), saturated sodium hydrogen carbonate solution (75
mL), and ethyl acetate (200 mL). After the solution was stirred
for 20 min, the phases were separated and the organic layer
washed in turn with a 3% citric acid solution (100 mL),
saturated sodium hydrogen carbonate solution (50 mL), and
brine (50 mL). The organic phase was dried over magnesium
sulfate, filtered, and concentrated in vacuo to give the crude
title compound as an off-white semicrystalline gum (1.8 g, w/w
assay by HPLC 45.7%; corrected yield 66%). A 1.7 g sample
of this material was recrystallized from n-hexane to give
compound 5 (0.51 g. 45%).
molar ratio 5:7, 15:1; ratio 5:3a , 21:1; ratio 5:6, 40:1. [R]20
)
D
29.6° (c ) 0.98, CHCl3); 1H-NMR (CDCl3, 500 MHz) δ 5.76 (bs,
1H), 4.18 (m, J ) 5.7, 6.0 Hz 1H), 4.00 (dd, J ) 2.66, 3.8 Hz,
1H), 3.57 (t, J ) 3.3 Hz, 1H), 3.28 (s, 3H), 3.10 (m, 1H), 2.88
(dd, J ) 2.6, 5.7 Hz, 1H), 2.24 (m, 1H), 2.11 (m, 1H), 2.01 (m,
1H), 1.69 (m, 1H), 1.66 (m, 1H), 1.56 (m, 1H), 1.25 (d, J ) 6.0
Hz, 3H), 0.88 (s, 9H), 0.09 (s, 3H), 0.007 (s, 3H); 13C-NMR
(CDCl3, 100.6 MHz) δ 213.30, 168.57, 84.09, 66.11, 60.88,
57.01, 49.46, 48.53, 33.75, 28.26, 25.72, 22.52, 19.07, 17.89,
-4.23, -5.07; IR (nujol mull) νmax 3202, 1759, 1718 cm-1; MS
(FAB/NBA) m/ z 356 [MH]+, 340, 324, 298, 181 (100), 156.
Anal. Calcd for C18H33NO4Si: C, 60.79; H, 9.37; N, 3.94.
Found: C, 60.73; H, 9.34; N, 4.06.
(3S,4R)-3-[(R)-1-[(ter t-Bu tyld im eth ylsilyl)oxy]eth yl]-4-
[(R)-6′-[(S)-2′-m eth oxy]-1′-oxocycloh exyl]a zetid in -2-on e
(5). Tin(IV) chloride (6.95 mL, 59.1 mmol) was added dropwise
to an efficiently stirred solution of (2S)-6 (2.65 g, 20.3 mmol)
in anhydrous chlorobenzene (75 mL) under nitrogen, main-
taining the temperature below -20 °C. A solution of 3a (5 g,
17.4 mmol) in anhydrous chlorobenzene (25 mL) was added
dropwise over 15 min to the reaction mixture, keeping the
temperature below -20 °C. The mixture was warmed to 0 °C
and N,N-diisopropylethylamine (7.88 mL, 45.2 mmol) dissolved
in anhydrous chlorobenzene (25 mL) was added dropwise over
20 min, maintaining the reaction temperature between 0 and
5 °C. The resulting yellow solution was stirred at 0-5 °C for
1 h and then poured onto a chilled (0-5 °C) mixture of a
saturated solution of NaHCO3 (150 mL) and a saturated
solution of of Rochelle’s salt (150 mL). The mixture was stirred
for 1 h, and the organic phase was separated. The aqueous
layer was extracted with ethyl acetate (150 mL). The organic
Ack n ow led gm en t. We thank Dr. Clive Meerholz for
helpful suggestions, Sean Lynn for providing the X-ray
data, and Alberto Zanoni for HPLC analysis.
Su p p or tin g In for m a tion Ava ila ble: Characterization
data for compounds 7, 13, and 14 (6 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
J O962095M