Synthesis of (1′R,3S,4S)-3-[1′-(tert-butyldimethylsilyloxy) ethyl]-4-(cyclopropylcarbonyloxy)azetidin-2-one

Article abstract of DOI:10.1002/ejoc.200600235

The novel carbapenem precursor 1e has been synthesized from L-threonine, cyclopropyl methyl ketone and benzhydrylamine (for the introduction of the azetidinone N-protecting group). Two independently prepared building blocks - sodium (2R,3R)-2,3-epoxybutyrate as a mixed salt with NaBr (2b) and N-(benzhydryl)aminomethyl cyclopropyl ketone (4e) - were coupled to give (2R,3R)-N-(benzhydryl)-N-(2-cyclopropyl-2-oxoethyl)-2,3-epoxybutyramide (8e). This key intermediate gave a regio- and stereoselective C3-C4 ring closure on LiHMDS treatment in THF at 0°C to yield (1′R,3S,4S)-4- cyclopropylcarbonyl-1-diphenylmethyl-3-(1-hydroxyethyl)azetidin-2-one (13e). N-Deprotection of 13e was performed by photochemical bromination and subsequent hydrolysis. The resulting (1′R,3S,4S)-4-(cyclopropylcarbonyl)-3-(1- hydroxyethyl)azetidin-2-one (23e) reacted in a Baeyer-Villiger oxidation with a total control of the regioselectivity (due to the poor migratory aptitude of the cyclopropyl group) to furnish (1′R,3S,4S)-3-(1-hydroxyethyl)-4- (cyclopropylcarbonyloxy)azetidin-2-one (24e), subsequent O-silylation achieving the total synthesis of 1e (title compound). Wiley-VCH Verlag GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejoc.200600235

FULL PAPER
DOI: 10.1002/ejoc.200600235
Synthesis of (1ЈR,3S,4S)-3-[1Ј-(tert-Butyldimethylsilyloxy)ethyl]-
4-(cyclopropylcarbonyloxy)azetidin-2-one
Mathieu Laurent,[a] Marcel Cérésiat,[b] and Jacqueline Marchand-Brynaert*[a]
Keywords: Azetidinone / Carbapenem precursor / Benzhydryl protecting group / Directed Baeyer–Villiger oxidation
The novel carbapenem precursor 1e has been synthesized
from -threonine, cyclopropyl methyl ketone and benzhy-
was performed by photochemical bromination and subse-
quent hydrolysis. The resulting (1ЈR,3S,4S)-4-(cyclopro-
L
drylamine (for the introduction of the azetidinone N-protect- pylcarbonyl)-3-(1-hydroxyethyl)azetidin-2-one (23e) reacted
ing group). Two independently prepared building blocks –
sodium (2R,3R)-2,3-epoxybutyrate as a mixed salt with NaBr
(2b) and N-(benzhydryl)aminomethyl cyclopropyl ketone
in a Baeyer–Villiger oxidation with a total control of the re-
gioselectivity (due to the poor migratory aptitude of the cy-
clopropyl group) to furnish (1ЈR,3S,4S)-3-(1-hydroxyethyl)-4-
(4e) – were coupled to give (2R,3R)-N-(benzhydryl)-N-(2-cy- (cyclopropylcarbonyloxy)azetidin-2-one (24e), subsequent
clopropyl-2-oxoethyl)-2,3-epoxybutyramide (8e). This key
intermediate gave a regio- and stereoselective C3–C4 ring
closure on LiHMDS treatment in THF at 0 °C to yield
O-silylation achieving the total synthesis of 1e (title com-
pound).
(1ЈR,3S,4S)-4-cyclopropylcarbonyl-1-diphenylmethyl-3-(1- (© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
hydroxyethyl)azetidin-2-one (13e). N-Deprotection of 13e
Germany, 2006)
synthesis of 1 are now available from the Japanese compa-
nies Kaneka, Nippon Soda and Takasago International.
Their respective strategies for the construction of the four-
membered heterocycle (azetidinone) rely on: (i) the [2+2]
cycloaddition of chlorosulfonyl isocyanate (CSI) to
(3R,1E)-3-(tert-butyldimethylsilyloxy)-1-(trimethylsilyloxy)-
butene,[3] (ii) CSI addition to (3R,1E)-3-(tert-butyldimethyl-
silyloxy)-1-(phenylthio)but-1-ene,[4] and (iii) the N1–C2 cy-
clization of (2R,3R)-2-(aminomethyl)-3-hydroxybutanoic
acid.[5] In both cycloaddition approaches the alkene chiral-
ity is provided by the microbial hydroxylation of methyl bu-
tanoate, while in the cyclization method the chiral precursor
is the result of catalytic hydrogenation of methyl 2-acetyl-
3-(benzoylamino)propanoate in the presence of a chiral Ru
catalyst.[6]
Introduction
(1ЈR,3S,4S)-4-Acetoxy-3-[1-(tert-butyldimethylsilyloxy)-
ethyl]azetidin-2-one (1, R = Me; Scheme 1) still remains the
most popular key intermediate for the synthesis of thiena-
mycin and carbapenem derivatives, and other novel antibio-
tics that might defeat bacterial resistance.[1]
Many other strategies (e.g., [2+2] cycloadditions of ke-
tenes to Schiff bases, [2+2] cycloadditions of ester enolates
to Schiff bases, N1–C4 cyclizations and C3–C4 cyclizations)
have been disclosed and extensively discussed,[2] but none
has been developed for the production of 1 on a large scale.
In this context, a communication from Hanessian et al. at-
tracted our attention:[7] azetidinone 1 (R = Ph) was pre-
pared by the C3–C4 cyclization of a 2,3-epoxybutyramide
precursor obtained from -threonine. The retrosynthetic
scheme (Scheme 1) was based on three key steps: (a) a
Baeyer–Villiger oxidation to introduce the C4 functionality,
performed preferably after cleavage of PG (N-deprotection),
(b) C3–C4 cyclization by intramolecular nucleophilic attack
on the epoxide, and (c) the coupling of an α-amino ketone
synthon onto (2R,3R)-2,3-epoxybutyric acid. In our opin-
Scheme 1. Retrosynthetic scheme: (a)–(c) see text.
A tremendous amount of effort has been devoted to the
search for efficient preparations of this chiral azetidinone 1,
with possible industrial developments.[2] Three commercial
[a] Université Catholique de Louvain, Unité de Chimie Organique
et Médicinale, Bâtiment Lavoisier,
Place Louis Pasteur 1, 1348 Louvain-la-Neuve, Belgium
Fax: +32-10-474168
E-mail: marchand@chim.ucl.ac.be
[b] Tessenderlo Chemie S.A./N.V.,
Avenue du Trône 130, 1050 Bruxelles, Belgium
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2006, 3755–3766
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3755

M. Laurent, M. Cérésiat, J. Marchand-Brynaert
FULL PAPER
ion, this route offers interesting features for potential devel- dialkylated anilines. Products 3a (R = Ph) and 3b (R = tBu)
opment (cheap chiral starting material, high stereocontrol, were recovered after precipitation of the hydrobromide salt
good yields, limited number of steps, smooth conditions, from hexane, filtration and concentration of the organic
…), provided that the N-protecting group can be modified. phase. Two different methods allowed the preparation of
Indeed, in the original paper, the authors made use of the synthons 4 [Equation (2)], working either in methanol with
p-methoxyphenyl (PMP) group, which requires suprasto- 2 equiv. of benzhydrylamine or in DMF with 1 equiv. of
ichiometric amounts of CAN (ceric ammonium nitrate) to benzhydrylamine and 1 equiv. of another added base (or-
be cleaved.[8] We therefore decided to revisit Hanessian’s ganic or inorganic). Compounds 4a (R = Ph) and 4b (R =
synthesis, but replacing the N-(p-methoxyphenyl) (PMP) tBu) were readily obtained and isolated by the first method,
substituent with an N-benzhydryl or N-(p,pЈ-dimethoxy)- because they precipitated in methanol, while benzhy-
benzhydryl group, which should enable deprotection by hy- drylamine hydrobromide remained soluble. Compounds 5a
drogenation or acidic hydrolysis.[9] What was expected to (R = Ph) and 5b (R = tBu) were prepared similarly [Equa-
be a trivial structural modification proved to bring major tion (3)]. The second method [Equation (2)] was used for
changes in the chemical reactivity and selectivity of the C3– the synthesis of 4c (R = o-MeOC6H4) with triethylamine as
C4 cyclization key step with it,[10] so the experimental con- a base and of 4d (R = iPr) and 4e (R = cyclo-Pr) with
ditions and the nature of the substituent R were adapted. potassium carbonate as a base and potassium iodide to ac-
Other surprises, however, also arose from the deprotection tivate the substrate. In these cases an aqueous workup was
step[11] and the Baeyer–Villiger oxidation.[12] Here we fully required to eliminate the salts. Lastly, the masked α-amino
describe our adventures in this azetidinone story and the aldehyde
6 was prepared as shown in Equation (4)
happy ending we were able to find for the synthesis of 1, (Scheme 2): benzhydryl bromide and 2 equiv. of 2-amino-
thanks to the discovery of a novel method of N-benzhydryl ethanal dimethyl acetal were allowed to react at 70 °C with-
deprotection and the use of the cyclopropyl substituent to out solvent, with the product being isolated after aqueous
direct the Baeyer–Villiger rearrangement (Scheme 1; R = workup and distillation.
cyclo-Pr and PG = CHPh2).
Results and Discussion
According to the retrosynthetic plan shown in Scheme 1,
the chiral building block required to start the synthesis of
azetidinone 1 is (2R,3R)-2,3-epoxybutyric acid (2a). This
acid is poorly stable and not easily prepared on a large
scale, so we adapted previous methods[13–16] and prepared
a mixed salt 2b with sodium carboxylate and sodium bro-
mide (1:1) (Scheme 5, Part A). -Threonine was thus treated
with KBr and NaNO2 in 2.5  H2SO4 to produce the corre-
sponding α-bromo acid derivative with retention of config-
uration.[17] This compound, extracted with ethyl acetate,
was then cyclized by treatment with 2 equiv. of NaOH in
water. Epoxide 2b was obtained directly by concentration
of the aqueous phase. This salt 2b could be stored at room
temperature without degradation; it was found by DTA
(differential thermal analysis) to be stable to 100 °C. As a
result of the SN2 mechanism, a single stereoisomer (2R,3R)
of epoxybutyrate was formed. The enantiomeric purity of
2b was monitored by GC after derivatization as the isopro-
pyl ester.[18]
Scheme 2. Synthesis of α-amino ketone reagents.
The second building block on the pathway to azetidinone
1 are N-protected (PG) α-amino ketones (3–6) (Scheme 1).
Such compounds were readily available from α-bromo
Different conditions were explored to perform the coup-
ketones and p-anisidine [PG = C6H4OMe, Scheme 2, Equa- ling reactions between epoxybutyrate 2b and amino ketones
tion (1)], benzhydrylamine [PG = CHPh2, Scheme 2, Equa- 3–6, with use of various activation reagents such as thionyl
tion (2)], or bis(p-methoxyphenyl)methylamine[19] [PG = chloride, oxalyl chloride, pivaloyl chloride, isobutyl chloro-
CH(C6H4OMe)2, Scheme 2, Equation (3)]. THF was the formate and 2,4,6-trichlorobenzoyl chloride. Our aim was
best solvent[20] for performing the substitution of Equa- to produce the acid chloride (or mixed anhydride) of 2b in
tion (1) with 2 equiv. of p-anisidine (1 equiv. acting as a situ, without proceeding through the unstable acid 2a. Oxal-
base); indeed, by the procedure of Tachdjian et al.[21] in yl chloride afforded the best results as coupling agent
diethyl ether we obtained a 75:25 mixture of mono- and (Table 1). Treatment of 2b with the p-anisidine derivatives
3756
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 3755–3766

Synthesis of a Novel Azetidin-2-one Carbapenem Precursor
FULL PAPER
3a and 3b gave moderate yields of epoxybutyramides 7a (R Two conformers are thus visible in solution at room tem-
= Ph) and 7b (R = tBu), bearing the PMP protecting group perature, and their structures were attributed by 2D
previously used by Hanessian et al.[7] (Entries 1–2). On the NOESY (nuclear Overhauser and exchange spectroscopy)
other hand, moderate to good yields were collected with experiments performed at 0 °C (Figure 1). The major rot-
the benzhydrylamine derivatives 4a to 4e (Entries 3–7), with amer showed the CH2COR moiety away from the epoxide.
epoxybutyramides 8a (R = Ph), 8b (R = tBu), 8c (R = o- Interestingly, in the solid state, 8b has been “frozen” in the
MeOC6H4), 8d (R = iPr) and 8e (R = cyclo-Pr) being puri- form of the minor conformer (X-ray structure[64]).
fied either by column chromatography on silica gel or by
crystallization. The preparation of compound 9 (R = H)
required two steps: the coupling of 2b and 6 to give the
intermediate 11 in 70% yield after chromatography (Fig-
ure 3) and acidic hydrolysis in a two-phase system (Entry 8).
The precursors 10a (R = Ph) and 10b (R = tBu), protected
with the bis(p-methoxyphenyl)methyl group, were prepared
similarly, but in low yields after purification (Entries 9–10).
Figure 1. Rotamers of 8b with NOESY correlations indicated.
The next step of our synthesis of 1 was the cyclization of
the epoxybutyramide precursors to azetidinones 12–15 un-
der basic conditions [Scheme 1, step (b)]: the carbanion
formed by abstraction of a hydrogen atom α to the COR
group could attack the epoxide moiety intramolecularly.
Both the nucleophilic and electrophilic parts of the reagents
being ambident functions, four cyclized products could be
formed in principle:[10] the azetidinones A (enolate C-alky-
lation at the C2 atom of the epoxide ring), the 5,6-dehy-
dromorpholin-3-ones B (enolate O-alkylation at the C2
atom of the epoxide ring), the 4,5,6,7-tetrahydro-4-azaoxe-
pin-5-ones C (enolate O-alkylation at the C3 atom of the
epoxide ring) and the pyrrolidin-2-ones D (enolate C-alky-
lation at the C2 atom of the epoxide ring). Experimentally,
three products, A, B and C, were isolated and carefully
identified, with ratios depending on the nature of the N-
protecting group (PG), the nature of the R substituent and
the experimental conditions (Table 3).
Table 1. Coupling to epoxybutyrate: i: 2b (1.2 equiv.), (COCl)2
(1.2 equiv.), THF, –5 °C, 2 h; ii: pyridine (3 equiv.), amine 3–6
(1 equiv.), 1 h at –5 °C and 2 h at 20 °C.
[a] Protected reagent 6 was used for the coupling, followed by acidic
hydrolysis (CHCl3, H2O, TFA, 3 h, 20 °C).
The NMR spectra of the epoxybutyramides recorded at
First of all we reproduced the synthesis of Hanessian et
25 °C showed interesting features: the PMP derivatives 7 al.,[7] with the same experimental conditions and starting
each appeared as a single rotamer giving one set of signals. material and, fortunately, found the same result (Entry 1):
The presence of two rotamers in rapid equilibrium was not treatment of epoxybutyramide 7a (PG = PMP and R = Ph)
considered as an alternative hypothesis because the spectra with potassium carbonate in DMF at high temperature fur-
did not show any splitting of the signals when recorded at nished the azetidinone 12a[7] as the only identified cyclized
lower temperature (CDCl3, –50 °C and acetone, –85 °C). product (isolated pure product: 76% yield). Similarly, pre-
On the other hand, the diarylmethyl derivatives 8–10 cursor 7b (PG = PMP and R = tBu) exclusively gave the
showed splitting of all the signals, characteristic of a slow- azetidinone 12b (isolated pure product: 81% yield) (En-
ing down of the rotation around the amide bond (Table 2). try 2).
1
Table 2. H NMR spectroscopic data for epoxybutyramides.
Compd. (Table 1)
Conformer ratio
δ(CH2)
JAB [Hz]
major minor
δ(CHAr2)
major minor
major
minor
7a[7]
7b[55]
8a
–
–
4.87:5.42
4.41:4.84
17.3
17.5
–
–
53:47
56:44
50:50
54:46
58:42
70:30
69:31
58:42
68:32
4.71:4.86
4.23:4.45
4.65:4.82
4.14:4.31
4.26:4.42
3.54:3.61
4.00:4.12
4.66:4.86
4.18:4.44
4.68:5.28
4.23:4.98
4.71:5.11
4.12:4.82
4.25:4.82
3.60
4.20:4.40
4.66:5.24
4.22:4.92
17.3
17.4
18.0
17.7
17.1
15.6
17.6
17.1
17.3
18.7
19.3
18.9
19.2
20.1
–
19.1
18.8
18.8
6.71
6.62
6.65
6.62
6.61
6.55
6.66
6.58
6.48
7.15
7.22
7.15
7.20
7.20
7.13
7.17
6.72
6.76
8b
8c
8d
8e
11
9
10a[56]
10b
Eur. J. Org. Chem. 2006, 3755–3766
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3757

M. Laurent, M. Cérésiat, J. Marchand-Brynaert
FULL PAPER
Table 3. Cyclization of epoxybutyramides.
[a] i: K2CO3, DMF, 100 °C, 18 h; ii: Li2CO3, DMF, 100 °C, 18 h; iii: LiHMDS, THF, 0 °C, 3 h. [b] Determined by 1H NMR spectroscopy
of the crude mixtures (rel. %).
The next experiments were somewhat disappointing. The was favoured by the presence of the tert-butyl substituent
replacement of the PMP protecting group by the benzhy- (steric effect), while the presence of the phenyl group in-
dryl group resulted in the formation of mixtures of cyclized creased the ratio of O-alkylation (electronic effect). Hence,
products when epoxybutyramides 8a (PG = Ph2CH and R extension of the mesomeric effect (R = o-MeOC6H4) led to
= Ph) and 8b (PG = Ph2CH and R = tBu) were treated 31% of azetidinone 13c (Entry 9) instead of 57% (R = Ph).
with K2CO3 in hot DMF (Entries 3 and 6). However, the The steric effect was also demonstrated further by replace-
amount of azetidinone 13 was higher with the R substituent ment of the tBu group with the smallest substituent, a hy-
being tBu (13b, 55%) rather than phenyl (13a, 16%). In the drogen atom. Under all tested conditions, treatment of ep-
case of the bis(p-methoxyphenyl)methyl protecting group, oxybutyramide 9 with a base gave O-cyclized products and
the results were in the same range [Entries 15 and 18; PG azetidinone 14 only as traces (Յ 2%; Entries 12–14; PG =
= (p-MeOC6H4)2CH and R = Ph or tBu]. At this stage we Ph2CH and R = H). With the isopropyl and cyclopropyl
speculated that the nature of the base and the counter-cat- substituents, on the other hand, high yields (85–88%) of
ion might significantly modify the enolate reactivity and the azetidinones 13d and 13e, respectively, were collected on
C/O-alkylation selectivity: a soft cation should favour O- treatment of the precursors 8d and 8e with LiHMDS in
alkylation while a hard cation should favour C-alky- THF (Entries 10 and 11; PG = CHPh2 and R = iPr or cy-
lation.[22] Indeed, by using lithium carbonate as a base in clo-Pr). Quantitative formation of azetidinone (13b), how-
hot DMF we increased the amounts of azetidinones: 34% ever, could be observed only when R was tert-butyl and PG
of 13a instead of 16% with K2CO3 (Entry 4; PG = Ph2CH benzhydryl (Entry 8).
and R = Ph), 62% of 13b instead of 55% (Entry 7; PG =
The structures of the cyclized products 12–15 (A), 16–18
1
Ph2CH and R = tBu), 30% of 15a instead of 21% [En- (B) and 19–21 (C) (Table 3) were assigned from the H and
try 16; PG = (p-MeOC6H4)2CH and R = Ph], and 65% of 13C NMR spectroscopic data and, in some cases, X-ray dif-
15b instead of 61% [Entry 19; PG = (p-MeOC6H4)2CH and fraction analyses of single crystals.[23–25] Three protons are
R = tBu]. This effect is even more pronounced with lithium characteristic in azetidinones 12–15: namely, Ha on the 1-
hexamethyldisilazide (LiHMDS) as a base in THF, a less hydroxyethyl chain, giving a signal (dq) at δ = 4.2–4.3 ppm,
polar solvent in which the ionic pairs are more intimate. Hb on the β-lactam ring (C3), giving a doublet of doublets
The amounts of azetidinones were 57% of 13a (Entry 5; PG at δ = 2.9–3.1 ppm, and Hc on the β-lactam ring (C4), giv-
= Ph2CH and R = Ph), 100% of 13b (Entry 8; PG = Ph2CH ing a doublet at δ = 4.5–5.1 ppm (Figure 2, Table 4). The
and R = tBu), 41% of 15a [Entry 17; PG = (p-MeOC6H4) coupling constant of 2–2.5 Hz between Hb and Hc is typi-
2CH and R = Ph], and 74% of 15b [Entry 20; PG = (p- cal of trans stereochemistry. The absolute configurations of
MeOC6H4)2CH and R = tBu]. Comparison between En- the chiral centres were inferred from the known chirality of
tries 5 and 8 and Entries 17 and 20 pointed to the effect of the starting material and the SN2i mechanism of epoxide
the R substituent on the enolate reactivity. The C-alkylation substitution (inversion of configuration) and were con-
3758
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 3755–3766

Synthesis of a Novel Azetidin-2-one Carbapenem Precursor
FULL PAPER
firmed by X-ray data.[25] In the 13C NMR spectrum, the β- chain giving the same pattern as the corresponding proton
lactam carbonyl signal is visible at δ = 164–167 ppm, and in azetidinones A (δ = 4.3–4.5 ppm), Hb on the morpholine
the ketone carbonyl signal at δ = 200–211 ppm, whilst the heterocycle giving a doublet at δ = 4.2–4.5 ppm, and the
IR spectra show the typical β-lactam stretching at 1740– olefinic proton Hc giving a singlet at δ = 5.4–6 ppm (doub-
1760 cm–1. Azetidinones 12–15 (Table 4) were obtained ex- let for R = H). The presence of a C=C double bond is
clusively as the trans stereoisomers and the side-products visible in the 13C NMR with signals at δ = 100–103 ppm
arising from the cis stereoisomers (Ϸ 10%), described by (C5) and δ = 130–150 ppm (C6). The stereochemical assign-
Kugelman et al.[26] (PG = PMP and R = C6H4X), were not ment, in agreement with the reaction mechanism, was con-
formed in our hands.
firmed by X-ray data.[23]
The tetrahydroazaoxepinones 19–21 (C) (Table 3) are
also each characterized by three protons (Figure 2, Table 6):
Ha on the heterocycle (δ = 4.4–4.8 ppm), Hb also on the
heterocycle (δ = 4.4–4.9 ppm) and the olefinic proton Hc (δ
= 5–6 ppm), and by the C=C double bond (C2 at δ = 130–
150 ppm and C3 at δ = 100–110 ppm). Here the X-ray dif-
fraction analysis was again useful for structural and stereo-
chemical assignment.[24]
Our results could be interpreted in terms of Baldwin’s
rules[27–29] and conformational (rotamers around the amide
bond) and configurational [(E)/(Z)-enolates] considerations.
Of the four possible cyclization processes: three are allowed
Figure 2. Atom numbering considered in Tables 4–6.
The dehydromorpholinones 16–18 (B) (Table 3) each by Baldwin’s rules (4-exo-tet, 6-exo-tet and 7-endo-tet)
show signals of three characteristic protons in the 1H NMR whilst one is disfavoured (5-endo-tet), and indeed, this last
spectrum (Figure 2, Table 5): Ha on the 1-hydroxyethyl process has not been observed experimentally. The cycliza-
Table 4. NMR spectroscopic data for azetidinones (Figure 2).
Compd. (Table 3) δ(Ha) (mult, J [Hz])
δ(Hb) (mult, J [Hz])
δ(Hc) (mult, J [Hz]) δ(Hd) (mult) δ(C2) δ(C3) δ(C4) δ(C5) δ(C6) δ(C7)
12a[7]
2b[55]
13a
13b
13c
4.33 (quint, 6.4)
4.26 (dq, 6.3, 6.9)
4.31 (dq, 5.2, 6.3)
4.21 (dq, 5.4, 6.4)
4.22 (m)
4.22 (dq, 5.4, 6.6)
4.26 (m)
4.26 (dq, 5.1, 6.2)
4.17 (m)
3.19 (dd, 2.4, 6.4)
3.05 (dd, 2.1, 6.9)
3.11 (dd, 2.5, 5.2)
2.89 (dd, 2.1, 6.4)
3.14 (dd, 2.5, 4.6)
2.89 (dd, 2.4, 5.4)
3.09 (dd, 2.4, 4.8)
3.08 (dd, 2.2, 5.1)
2.87 (dd, 2.0, 5.9)
3.07 (m)
2.87 (dd, 2.1, 6.3)
2.84 (dd, 2.5, 5.6)
2.95 (dd, 2.5, 5.1)
3.23 (ddd, 2.5, 4.0, 6.3)
3.21 (m, 2.4, 6.3)
3.11 (m)
5.54 (d, 2.4)
4.97 (d, 2.1)
5.07 (d, 2.3)
4.52 (d, 2.1)
5.14 (d, 2.5)
4.34 (d, 2.4)
4.41 (d, 2.4)
5.07 (d, 2.2)
4.50 (d, 2.0)
5.21 (d, 2.0)
4.69 (d, 2.1)
4.60 (d, 2.5)
4.62 (d, 2.5)
5.07 (d, 2.5)
4.58 (d, 2.1)
4.42 (d, 3.0)
4.45 (d, 2.7)
5.84 (s)
–
–
not determined
163.6
167.1
166.9
167.4
166.8
167.1
62.6
62.3
61.9
62.1
61.2
61.5
55.5
55.3
55.3
58.6
57.9
59.9
66.6 211.0
–
65.2 197.6 62.5
66.2 212.7 62.8
64.3 199.9 62.5
65.1 211.5 62.5
64.4 208.0 62.1
5.59 (s)
5.69 (s)
5.88 (s)
5.85 (s)
5.95 (s)
5.82 (s)
5.57 (s)
13d
13e
15a
15b
22a
22b
22d
22e
not determined
167.1
61.7
55.3
66.2 213.0 55.8
4.29 (m)
–
–
–
–
–
–
–
–
–
–
–
–
–
–
not determined
4.26 (dq, 6.3, 6.6)
4.28 (dq, 5.6, 6.3)
4.22 (dq, 5.1, 5.8)
4.32 (quint, 6.3)
4.24 (quint, 6.3)
4.29 (quint, 6.0)
4.33 (dq, 4.4, 6.0) 3.22 (ddd, 1.6, 2.7, 4.4)
4.24 (quint, 6.0)
4.09 (dq, 5.7, 6.3)
4.28 (dq, 3.5, 6.3) 3.21 (ddd, 1.1, 2.5, 3.5)
166.7
166.4
166.7
167.6
169.0
168.4
167.2
167.2
166.8
168.6
168.3
166.4
166.4
60.8
60.6
60.7
63.7
64.1
63.4
63.7
63.9
64.3
63.4
64.0
65.0
63.9
57.2
59.1
60.5
54.5
52.6
55.3
57.3
75.9
76.5
49.9
50.0
75.2
75.0
66.2 216.0 87.4
65.0 213.9 87.6
64.5 209.7 87.6
23a[57]
23b
23d
23e
66.4 196.6
65.9 213.0
64.6 212.3
65.0 207.6
65.2 177.6
66.0 175.0
64.0 170.5
64.5 170.8
64.0 177.3
65.0 175.0
–
–
–
–
–
–
–
–
–
–
24d
24e
3.24 (dd, 1.2, 6.0)
3.15 (dd, 1.5, 5.7)
5.90 (s)
25b[58]
25d
1d
4.18 (d, 2.5)
4.29 (d, 2.4)
5.82 (s)
4.34 (dq, 3.3, 6.3)
4.22 (dq, 3.6, 6.3)
4.22 (dq, 3.6, 6.3)
3.31 (dd, 2.4, 3.3)
3.19 (dd, 1.5, 2.4)
3.19 (dd, 1.2, 3.6)
1e
5.83 (s)
Table 5. NMR spectroscopic data for dehydromorpholinones (Figure 2).
Compd. (Table 3)
δ(Ha) (mult, J [Hz])
δ(Hb) (mult, J [Hz]) δ(Hc) (mult, J [Hz]) δ(Hd) (mult) δ(C2) δ(C3) δ(C5) δ(C6) δ(C7) δ(C8)
16a
16b
17
18a
18b
4.50 (m)
4.37 (ddq, 4.1, 6.4, 6.6)
4.52 (dq, 2.7, 6.6)
4.45 (m)
4.50 (m)
4.21 (d, 4.1)
4.48 (d, 2.7)
4.45 (m)
6.07 (s)
5.30 (s)
5.88 (d, 6.9)
6.05 (s)
7.10 (s)
7.00 (s)
7.15 (s)
6.96 (s)
6.90 (s)
79.3
79.9
76.7
163.6 103.7 139.7
164.2 101.8 151.1
171.6 100.7 132.7
not determined
66.7
67.3
72.8
59.7
60.2
61.7
4.37 (m)
4.26 (d, 4.1)
5.45 (s)
79.1
166.5 100.8 150.5
66.6
54.8
Eur. J. Org. Chem. 2006, 3755–3766
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3759

M. Laurent, M. Cérésiat, J. Marchand-Brynaert
FULL PAPER
Table 6. NMR spectroscopic data for tetrahydroazaoxepinones (Figure 2).
Compd. (Table 3) δ(Ha) (mult, J [Hz]) δ(Hb) (mult, J [Hz]) δ(Hc) (mult, J [Hz]) δ(Hd) (mult) δ(C2) δ(C3) δ(C5) δ(C6) δ(C7) δ(C8)
19a
19b
19c
19e
20
4.70 (dd, 4.0, 4.2)
4.57 (dd, 4.6, 5.2)
4.88 (t, 5.3)
4.54 (m)
4.40 (d, 4.0)
4.66 (br. s)
4.85 (dq, 4.2, 6.5)
4.69 (dq, 4.6, 6.3)
4.99 (quint., 5.9)
4.54 (m)
4.36 (dq, 4.0, 6.5)
4.80 (dq, 6.4)
5.72 (s)
5.05 (s)
6.17 (s)
4.99 (s)
6.18 (d, 4.4)
5.71 (s)
7.19 (s)
7.09 (s)
7.17 (s)
7.12 (s)
7.01 (s)
7.12 (s)
6.98 (s)
145.1 104.6 171.8
155.8 102.0 171.4
143.7 110.8 171.8
80.7
81.5
83.3
79.5
79.5
71.5
70.4
70.4
71.2
66.8
62.2
61.6
61.9
61.6
59.3
146.8
99.3
171.3
130.0 107.4 163.6
21a
21b
not determined
not determined
4.55 (t, 5.1)
4.68 (dq, 5.1, 6.4)
5.07 (s)
tion could occur from only one of the two possible con- be stable: the mixture of all cyclized products was never
formers around the N–C(O) amide bond (see Figure 1) in regenerated. Our results thus solely reflect the ratios of (E)-/
the epoxybutyramide precursors: namely the N conformer. (Z)-enolates initially formed and their possibility to equili-
Since the enolate stereochemistry [(E) or (Z)] should influ- brate or not. Surprisingly, the question of selectivity for in-
ence the selectivity of ring closure, the nature of the protect- tramolecular cyclization of epoxyenolate derivatives was
ing group (PG) plays a crucial role. An (E)-enolate could poorly documented in previous literature.[30–31] Crotti et
only provide a four-membered heterocycle while a (Z)-enol- al.[32–33] recently studied the effect of the chain length sepa-
ate could cyclize to six-, seven- but also four-membered het- rating a terminal epoxide from a benzoyl moiety on the
erocycles (Scheme 3). With PG = PMP, only one isomer of product distribution under basic conditions: the 4,5- and
the enolate [(E)-enolate] is obtained and the 4-exo-tet pro- 6,7-epoxides gave C-alkylation products (3-exo-tet, 5-exo-
cess was observed in all cases, while with the nonconjugated tet, 6-endo-tet), while the 5,6-epoxides furnished mainly O-
and bulkier benzhydryl group, the three allowed processes alkylation products (6-exo-tet, 7-endo-tet). In only one case
were in competition. In this latter case, the (E)-enolate is (bicyclic epoxide) has the disfavoured 5-endo-tet cyclization
probably less stable than its (Z) isomer, due to steric inter- been observed.[33] Huang[34] mentioned this type of cycliza-
actions between the PG and R groups. Thus, under revers- tion as a special case when the epoxide was substituted with
ible conditions of enolate formation (K2CO3, DMF, ∆), the a phenyl group.
reaction should proceed mostly via the (Z)-enolate, giving
With azetidinones 13 (PG = Ph2CH) and 15 [PG = (p-
three kinds of cyclized products. On the other hand, under MeOC6H4)2CH] now available, we examined the N-depro-
nonreversible conditions of enolate formation (LiHMDS, tection reaction under standard conditions (Table 7).[35] No
THF, 0 °C), the (E)-enolate should be formed more rapidly, reaction was observed under catalytic hydrogenation condi-
giving the expected azetidinone as the major (or unique) tions for the precursor 13b (PG = Ph2CH and R = tBu),
product. The (E)/(Z) ratios of the enolates and the reactivit- nor for the precursor 13a (PG = Ph2CH and R = Ph) (En-
ies of the ionic pairs are obviously also influenced by the tries 1–2), but in this case, when the conditions were forced,
nature of solvent and the temperature, but these parameters reduction of the benzoyl moiety occurred without concomi-
were chosen according to the base (mainly for solubility tant cleavage of the benzhydryl group. Similarly, we were
reasons). Lastly, we performed control experiments to en- unable to deprotect the precursor 15b [PG = (p-MeOC6-
sure that the cyclization step was nonreversible: each cy- H4)2CH and R = tBu] under various sets of acidic condi-
clized product (13b, 16b, 19b), placed individually under ba- tions (Entry 3). Prolonged reaction times with formic or tri-
sic conditions, (Li2CO3, DMF, 100 °C, 2 d) was found to fluoroacetic acids resulted in the acylation of the side-chain
Scheme 3. Cyclization processes (the process is named exo when the C–C bond of the epoxide is outside the formed cycle and endo when
the C–C bond of the epoxide is inside the formed cycle).
3760
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 3755–3766

Synthesis of a Novel Azetidin-2-one Carbapenem Precursor
FULL PAPER
hydroxy group, but without concomitant cleavage of the
bis(p-methoxyphenyl)methyl substituent.
Table 7. Azetidinone N-deprotection.
Scheme 4. Postulated routes for the azetidinone degradation.
These failures prompted us to seek a new deprotection
method compatible with the azetidinone stability.[11] We de-
cided to exploit the proradical character of the benzhydryl
motif. Inspired by some precedents involving benzylic cleav-
ages,[36–40] we treated the β-lactam 13b with N-bromosuc-
cinimide (NBS) in the presence of azobis(isobutyronitrile)
(AIBN) as a catalyst (CCl4 or PhCl, 80–120 °C). At high
temperatures, 13b indeed reacted, but furnished a complex
mixture of products (Scheme 4) resulting from azetidinone
ring opening (1H NMR analysis). At room temperature,
13b was slowly transformed to give a less complex mixture
in which the tetrahydrofuran derivative E (Scheme 4) was
was identified by 1H and 13C NMR spectroscopy: the char-
acteristic pattern of the Ha, Hb and Hc azetidinone protons
(Figure 2) is still present, but the benzhydryl proton Hd has
disappeared and two broad singlets attributable to OH pro-
tons are visible at δ = 2.50 and 5.12 ppm (hydroxyethyl
chain and benzhydryl group, respectively), whilst the signal
of the benzhydryl C7 signal at δ = 62.8 ppm has been re-
placed by a signal at δ = 87.4 ppm (benzhydrol carbon
atom). An X-ray diffraction analysis[25] confirmed our
structural attribution. The remarkable stability of the N-
acyl hemiaminal structure of 22b was an unexpected result
that we tentatively attributed to the involvement of the
benzhydrol hydroxy group in an H-bonding network.[41–43]
Thus, intermediate 22b was not spontaneously hydrolysed
in the biphasic reactive medium (CH2Cl2/H2O and HBr
formed during the radical bromination); this hydrolysis re-
quired subsequent treatment with p-toluenesulfonic acid
(PTSA) in wet acetone or acetonitrile. The N-unsubstituted
azetidinone 23b was recovered in high yield (88% for the
two steps of the benzhydryl deprotection; Table 7, Entry 4),
together with benzophenone (Ͼ 90% yield), which could be
recycled for the preparation of benzhydrylamine.
1
identified by H or 13C NMR analysis as the main product
[H2: δ = 1.44 (dd) ppm; H3: δ = 2.61 (dd) ppm; H4: δ =
3.87 (dq) ppm; H5: δ = 0.67 (d) ppm; C1: δ = 102.0 ppm;
C2: δ = 25.3 ppm; C3: δ = 56.2 ppm; C4: δ = 67.9 ppm; C7:
δ = 177.0 ppm]. The first radical formed by H abstraction
from the benzhydryl group could suffer β-scission of the
N1–C4 bond. This second radical could abstract a proton
and the resulting γ-hydroxy ketone should readily cyclize to
a hemiketal. In fact, at 20 °C, AIBN did not play any role;
the results were the same in the presence or absence of the
catalyst, provided that the reaction was performed under
light!
We then further considered photoactivation of the bro-
mination reaction of 13b. Under light, however, the ex-
pected product F appeared to be unstable (Scheme 4). For-
tunately, though, by working in a two-phase system, this
intermediate could be quenched by water to furnish the
stable β-lactam 22b quantitatively (Figure 3). We also found
that the addition of a catalytic amount of bromine ac-
celerated the photochemical process (Br2 is more sensitive
to light-induced homolytic cleavage than NBS). Accord-
ingly, the selected experimental conditions to transform 13b
Figure 3. Synthetic intermediates 11 and 22a–e.
Our two-step procedure for N-benzhydryl cleavage on β-
into 22b were as follows: NBS (1.1 equiv.), Br2 (0.05 equiv.), lactam cores was successfully applied to compounds 13a,
CH2Cl2/H2O (2:1), 20 °C, with stirring overnight under the 13d and 13e to furnish the NH-azetidinones 23a, 23d and
irradiation by the fume hood lamp. The benzhydrol deriva- 23e, respectively, in 85–95% yields (Table 7, Entries 5–7);
tive 22b was recovered in the organic phase. This compound their NMR structural assignments are summarized in
Eur. J. Org. Chem. 2006, 3755–3766
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3761

M. Laurent, M. Cérésiat, J. Marchand-Brynaert
FULL PAPER
Table 4. The cleavage of N-benzhydryl protecting groups azetidinyl Ͼ cyclo-Pr = Ph. The particular effect of the cy-
from aziridines, via the N-benzhydrol intermediates, has re- clopropyl group in relation to the other sterically hindered
cently been explored by Wulff et al.[44]
alkyl groups could be the result of the electronic properties
The last step of our strategy was the Baeyer–Villiger (B– of the exocyclic C–C bond of the cyclopropane ring and the
V) oxidation[45–46] of precursors 23, assuming that the re- reduced number of C–H bonds involved in the hyperconju-
quired regioisomer 24 (Table 8) would be favoured by the gative stabilization of the positive charge. Stereoelectronic
ability of the azetidinyl moiety to stabilize a partial positive effects most probably are not operating in this case.
charge created at C4. Indeed, as a general rule in B–V re-
The continuation of the synthesis towards O-protected
arrangements, it has been established that the more elec- azetidinones 1 (Scheme 1) was rather straightforward: treat-
tron-rich substituent should migrate preferentially. Involve- ment of 24d–e with tert-butyldimethylsilyl chloride
ment of stereoelectronic effects has also been demon- (TBDMSCl) and imidazole quantitatively furnished com-
strated,[47–48] however, meaning that the preference for mi- pounds 1d–e (see Table 4 for NMR spectroscopic data).
gration of one group can in some cases be overturned in (1ЈR,3S,4S)-3-[1-(tert-Butyldimethylsilyloxy)ethyl]-4-(cyclo-
favour of the other one, with an intrinsically lower mi- propylcarbonyloxy)azetidin-2-one (1e, R = cyclo-Pr) could
gration aptitude, by appropriate orbital interactions with be directly engaged in the further synthesis of carba-
the peroxocarboxyl leaving group. The migratory aptitude penems,[12] or transformed into the commercially available
of the azetidinyl group had not previously been systemati- acetoxy derivative 1 (R = Me, Scheme 1) by substitution
cally studied: in three cases – 4-acetylazetidin-2-one,[49] with acetate.[12]
4-benzoylazetidin-2-one,[50–52] and 4-formylazetidin-2-
one[53] – it was shown that the azetidinyl group migrated
preferentially over methyl, phenyl, and even hydrogen,
respectively.
Conclusions
At the end of this story, we are able to propose a novel
synthesis of the key carbapenem intermediate 1, by starting
Table 8. Baeyer–Villiger oxidation: i: mCPBA, CH2Cl2, 20 °C, 18 h.
from -threonine (Scheme 5, part A) and cyclopropyl
methyl ketone (part B) and using benzhydrylamine for the
introduction of the azetidinone N-protecting group. The
key step is the C3–C4 ring closure of the epoxybutyramide
precursor, regio- and stereocontrolled thanks to appropriate
choice of the experimental conditions and the steric effects
of the cyclopropyl substituent, similar to those of tert-butyl
(part C). A photohalogenation/hydrolysis sequence allowed
the smooth removal of the N-benzhydryl protecting group
Treatment of 4-benzoylazetidin-2-one 23a with m-chloro- on the azetidinone and the recovery of benzophenone
perbenzoic acid (mCPBA) in dichloromethane (20 °C, 18 h) (economy of atoms). The required C4 substitution was
quantitatively produced the corresponding 4-benzoyloxyaz- achieved through the B–V oxidative rearrangement with to-
etidin-2-one 24a (Table 8, Entry 1), as described by Hanes- tal regiocontrol, thanks to the poor migratory aptitude of
sian.[7] Under similar conditions, however, the tert-butylcar- the cyclopropyl group. The cyclopropyl substituent (R =
bonyl precursor 23b gave complete inversion of regioselec- cyclo-Pr in Scheme 1) thus represents the best compromise
tivity with the formation of 25b as the unique product of for control of both the cyclization and the B–V reactions:
B–V oxidative rearrangement, with the tert-butyl group mi- indeed this group provides strong steric/conformational ef-
grating preferentially over the azetidinyl group (Entry 2). fects, but poor electron-donating effects.
The isomers 24 and 25 could be easily distinguished in the
Practically, two building blocks were independently pre-
1H NMR spectrum thanks to the H4 signal: δ = 5.9 ppm pared: sodium (2R,3R)-2,3-epoxybutyrate (2b) and (benzhy-
for compound 24a and δ = 4.2 ppm for compound 25b drylamino)methyl cyclopropyl ketone (4e), obtained in 69%
(Table 4); the corresponding carbon C4 signals are also typ- and 81% yields, respectively. Coupling of the two buildings
ical in the 13C NMR spectrum: δ = 76 ppm (24) vs. δ = blocks, cyclization to the azetidinone, N-deprotection, B–V
50 ppm (25). As a matter of fact, the isopropyl derivative oxidation and O-silylation were performed in 47% overall
23d represents an intermediate situation producing an equi- yield (Scheme 5). Attempts to change the chronological or-
molar mixture of the regioisomers (24d + 25d, Entry 3) un- der of these reaction steps were not successful. The O-si-
der B–V conditions. Finally, the cyclopropyl derivative ful- lylation could be performed on azetidinones 13, but only
filled our expectation! Thanks to the increased electronega- under very harsh conditions (LiHMDS, 18-crown-6,
tivity of the cyclopropane carbon atoms (partial sp2 charac- 10 equiv. of TBDMSCl), due to the steric effect of the N-
ter),[54] the required regioselectivity was restored and the B– protecting group. This was also performed in a one-pot pro-
V oxidative rearrangement of 23e exclusively yielded the 4- cess, during workup of the cyclization step with LiHMDS.
(cyclopropylcarbonyloxy)azetidin-2-one 24e. Thus, in the The B–V oxidation, however, did not work at all in the pres-
series of ketones 23, the experimentally determined order ence of the bulky N-benzhydryl protecting group: neither
of group migration under B–V conditions is tBu Ͼ iPr = 13 (free OH), nor the O-silylated derivatives reacted with
3762
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 3755–3766

Synthesis of a Novel Azetidin-2-one Carbapenem Precursor
FULL PAPER
Scheme 5. Part A: i: NaNO2, KBr, H2SO4/H2O, –10 °C to 20 °C, 3 h; ii: NaOH, H2O, –10 °C to 20 °C, 18 h. Part B: i: Br2, MeOH, 0 °C
to 20 °C; ii: K2CO3, KI, Ph2CHNH2, DMF, 20 °C. Part C: i: (COCl)2, THF, –5 °C, 2 h, then pyridine, 2 h, 20 °C; ii: LiHMDS, THF,
0 °C, 3 h; iii: NBS, Br2 (catal.), CH2Cl2/H2O, hν, 20 °C, 15 h, then PTSA, acetone/H2O, 20 °C, 18 h; iv: mCPBA, CH2Cl2, 20 °C, 18 h; v:
TBDMSCl, imidazole, DMF, 20 °C, 48 h.
performed on commercial plates (aluminium sheets coated with sil-
ica Merck 5179, 250 mesh, with fluorescent indicator 60F 254,
0.25 mm thickness). Products were revealed with suitable indicators
such as phosphomolybdic acid (10% w/v in ethanol) and I2 vapour
or under a UV lamp. Column chromatography was carried out with
mCPBA, so N-deprotection has to precede B–V rearrange-
ment and O-protection.
Several advantages of our revisited Hanessian synthesis
could be pointed out: (1) the use of a stable form of the
chiral starting material (2b, mixed salt with NaBr), (2) the
use of a novel N-protecting group, deprotection of which
was not destructive and did not use cerium salt, (3) the
high regio- and stereoselectivities of the two key steps of
the reaction sequence, and (4) the introduction of the O-
1
silica gel 60 (70–230 mesh ASTM) supplied by Merck. H and 13C
NMR spectra were obtained with Varian Gemini 200 (at 200 MHz
for proton and 50 MHz for carbon), Varian Gemini 300 (at
300 MHz for proton and 75 MHz for carbon), and Bruker AM 500
spectrometers (at 500 MHz for proton and 125 MHz for carbon).
tert-butyldimethylsilyl group in the last step of the sequence Chemical shifts are reported in ppm (δ) downfield from tetrameth-
ylsilane (TMS) in deuterated chloroform or acetone, and DSS (2,2-
dimethyl-2-silapenta-5-sulfonate) in deuterated water. Multiplicity
is reported as follows: singlet (s), broad singlet (brs), doublet (d),
triplet (t), quadruplet (q), multiplet (m). Coupling constants (abso-
lute values) are expressed in Hertz (Hz). IR spectra were obtained
with a BIO-RAD FTS 135 instrument. Only the most significant
adsorption bands are reported. Mass spectra were recorded with a
Finnigan MAT TSQ-70 instrument. High-resolution mass spectra
(HRMS) were obtained from the University of Mons, Belgium
(Prof. R. Flammang). Microanalyses (EA) were performed at the
(TBDMSCl is an expensive reagent) because our conditions
of N-benzhydryl cleavage are compatible with a free OH
group (which was not the case for PMP cleavage with
CAN).
The scaling up of the synthesis of 1e for potential indus-
trial application is under investigation. Moreover, develop-
ments could be expected in fields other than azetidinone
chemistry, with the use of benzhydryl as an N-protecting
group for amides and lactams, and the use of cyclopropyl as
a directing substituent in Baeyer–Villiger rearrangements. Christopher Ingold Laboratories of University College, London
(Dr. A. Stones). Melting points were determined with an Electro-
thermal microscope and are uncorrected. Rotations were measured
with a Perkin–Elmer 241 MC polarimeter fitted with a sodium
lamp (concentrations expressed in g/100 mL). Some compounds
Experimental Section
cited in this paper have been described previously.[10–12] All com-
pounds are described in Tables 4–6 (NMR spectroscopic data), and
Table S1 (Supporting Information). The total synthesis of the title
compound (1e, Scheme 5) is described fully; the other derivatives
were similarly prepared.
General: Reagents and solvents were purchased from Acros, Ald-
rich, Fluka or Rocc. Technical quality solvents were used as eluents
in column chromatography and for liquid-liquid extractions. In all
other cases P.A. quality solvents were used. Reactions that required
anhydrous conditions were performed in glassware flamed under
vacuum and cooled in argon. Tetrahydrofuran (THF) was dried
with sodium in the presence of benzophenone and then distilled.
Dichloromethane was dried by heating at reflux in the presence of
calcium hydride and then distilled. DMF and acetonitrile were
dried with molecular sieves (4 Å). Thin layer chromatography was
Bromomethyl Cyclopropyl Ketone:[59] Br2 (11.3 mL, 220 mmol,
1 equiv.) was added in one portion to a cold solution (–5 °C) of
cyclopropyl methyl ketone (20.7 mL, 220 mmol, 1 equiv.) in freshly
distilled methanol (100 mL). After 1 h of stirring, water (50 mL)
Eur. J. Org. Chem. 2006, 3755–3766
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3763

M. Laurent, M. Cérésiat, J. Marchand-Brynaert
FULL PAPER
was added and the mixture was allowed to warm to 20 °C. After
stirring overnight, the solution was discoloured. After addition of
water (150 mL), the mixture was extracted with diethyl ether
(4×50 mL), and the organic phase was washed with NaHCO3
(5%), dried with MgSO4 and concentrated under reduced pressure
to furnish the crude product (37.5 g, yellow liquid; contamination
added dropwise at –5 °C to a suspension of finely powdered 2b
(3.36 g, 14.8 mmol, 1.2 equiv.) in dry THF (50 mL) (evolution of
gas occurred). After 2 h, pyridine (2.5 mL, 30.9 mmol, 2.5 equiv.)
and then amine 4e (3.09 g, 11.7 mmol, 1 equiv.) were added. After
1 h of stirring at –5 °C, the mixture was allowed to warm to 20 °C,
and stirred for 2 h. After dilution with ethyl acetate (50 mL), wash-
ing with NaHCO3 (5%, 3×30 mL), drying with MgSO4 and con-
centration, crude 8e was recovered. Purification by column
chromatography on silica gel (cyclohexane/ethyl acetate, 7: 3) gave
1
with about 10% of bromomethyl 3-bromopropyl ketone[60] by H
NMR analysis). Distillation (50 °C, 6 ×10–3 mbar) afforded pure
bromomethyl cyclopropyl ketone (29.5 g, 82% yield) as a colourless
1
liquid; d = 1.558 gmL–1. H NMR (300 Mz, CDCl3): δ = 1.13 (m, pure 8e (3.4 g, 83% yield) as a white solid (Table S1). 1H NMR
2 H), 1.22 (m, 2 H), 2.32 (tt, J = 4.8, 7.8 Hz, 1 H), 4.06 (s, 2 H)
(300 MHz, CDCL3, two rotamers in 58.5:41.5 ratio): δ = 0.6–0.8
(m, 4 H), 1.33 (d, J = 5.4 Hz, 3 H, minor isomer), 1.40 (d, J =
5.7 Hz, 3 H, major isomer), 1.5–1.7 (m, 1 H), 3.25 (m, 1 H), 3.54
(d, J = 4.8 Hz, 1 H, minor isomer), 3.59 (d, J = 4.5 Hz, 1 H, major
isomer), 4.25 (d, J = 20.1 Hz, 1 H, minor isomer), 4.26 (d, J =
17.1 Hz, 1 H, major isomer), 4.42 (d, J = 17.1 Hz, 1 H, major
isomer), 4.82 (d, J = 20.1 Hz, 1 H, minor isomer), 6.61 (s, 1 H,
major isomer), 7.15–7.35 (m + s, 10 H + 1 H, minor isomer) ppm.
13C NMR (75 MHz, CDCl3): δ = 10.7, 10.8, 10.9, 11.1, 13.9, 14.5,
17.6, 17.8, 53.5, 53.7, 53.8, 54.5, 55.2, 60.8, 63.9, 127.3–129.2,
ppm. 13C NMR (50 MHz, CDCl3): δ = 12.1, 18.4, 37.7, 201.7 ppm.
IR (film): ν = 3012, 2940, 1700, 1383, 1192, 1063, 895 cm–1. EIMS:
˜
m/z (%) = 163.8 (7) [81Br, M+·], 161.9 (7) [79Br, M+·], 153 (37), 122.8
(4), 120.9 (4), 94.7 (3), 92.9 (3), 68.9 (100).
(Benzhydrylamino)methyl Cyclopropyl Ketone (4e): Bromomethyl
cyclopropyl ketone (4.5 g, 28 mmol, 1 equiv.) was added at 90 °C
to a suspension of finely powdered K2CO3 (4.15 g, 30 mmol,
1.2 equiv.), KI (4.95 g, 30 mmol, 1.2 equiv.) and benzhydrylamine
(4.78 g, 27 mmol, 0.98 equiv.) in dry DMF (100 mL). This suspen-
sion was heated for 4 h and then filtered. The filtrate was diluted
with ethyl acetate (100 mL), washed with brine (3×50 mL), dried
with MgSO4 and concentrated under vacuum and the solid residue
(7.08 g, 98.8% yield) was used without purification in the next step.
A pure sample could be obtained by crystallization from methanol;
138.3, 138.5, 167.4, 167.9, 202.6, 203.6 ppm. IR (KBr): ν = 3062,
˜
3028, 3007, 2930, 1715, 1665, 1449, 1387, 1075, 729, 700 cm–1
.
CIMS: m/z (%) = 350.0 (4) [M+H+], 279.1 (12), 267.9 (8), 204.9
(20), 187.0 (16), 166.9 (56) [Ph2CH+], 153.9 (61); 60.9 (100).
(1ЈR,3S,4S)-1-Diphenylmethyl-3-(1-hydroxyethyl)-4-(cyclopropyl-
carbonyl)azetidin-2-one (13e): LiHMDS–diethyl ether[61] (3.10 g,
12.9 mmol, 1.2 equiv.) was added at 0 °C under argon to a solution
of 8e (3.72 g, 10.6 mmol, 1 equiv.) in dry THF (90 mL). After
warming up to 20 °C and 3 h of stirring, the reaction was stopped
by addition of HCl (1 , 10 mL). After dilution with ethyl acetate
(50 mL), the solution was washed with NaHCO3 (5%, 3×50 mL),
dried with MgSO4 and concentrated under vacuum to furnish
crude 13e (3.78 g, 100% yield). Purification by column chromatog-
raphy on silica gel (cyclohexane/ethyl acetate, 10:1) gave pure 13e
(3.04 g, 82% yield) as a white solid (Table S1). 1H NMR (300 MHz,
CDCl3): δ = 0.65–0.85 (m, 4 H), 1.29 (d, J = 6.7 Hz, 3 H), 1.81 (tt,
J = 4.5, 7.5 Hz, 1 H), 3.02 (br. s, 1 H), 3.09 (dd, J = 2.4, 4.8 Hz, 1
H), 4.26 (m, 1 H), 4.41 (d, J = 2.4 Hz, 1 H), 5.95 (s, 1 H), 7.20–
7.40 (m, 10 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 11.6, 12.3,
17.5, 21.8, 59.9, 61.5, 62.1, 64.4, 127.5–128.5, 138.0, 138.4, 167.1,
1
m.p. 82.5–84.0 °C (orange crystals). H NMR (200 Mz, CDCl3): δ
= 0.88 (m, 2 H), 1.06 (m, 2 H), 1.82 (tt, J = 4.5, 7.8 Hz, 1 H), 3.66
(s, 2 H), 4.81 (s, 1 H), 7.1–7.4 (m, 10 H) ppm. 13C NMR (50 MHz,
CDCl3): δ = 10.7, 18.8, 58.1, 67.0, 126.9, 127.4, 128.5, 143.6,
208.6 ppm. IR (KBr): ν = 3326, 3083, 3060, 3025, 3006, 2880,
˜
2786,1702, 1451, 1391, 1069, 698 cm–1. EIMS: m/z (%) = 265.1 (9)
[M+·], 181.9 (75) [Ph2CHNH+], 167.0 (38) [Ph2CH+], 104.9 (100),
76.9 (63), 68.9 (47). HR-CIMS: calcd. for C18H20NO 266.1545,
found 266.1544 [M+H+].
(2S,3R)-2-Bromo-3-hydroxybutyric Acid:[17] A solution of NaNO2
(9.38 g, 136 mmol, 1.6 equiv.) in water (30 mL) was added drop-
wise, over 90 min, to a cold (–12 °C) solution of -threonine (10 g,
84 mmol, 1 equiv.) and KBr (15.48 g, 130 mmol, 1.5 equiv.) in
H2SO4 (2.5 , 170 mL). The solution was allowed to warm to 20 °C
and stirred for 3 h and the aqueous solution was extracted with
ethyl acetate (6×50 mL). The organic phase was washed with brine,
dried with MgSO4 and concentrated under vacuum to furnish
208.0 ppm. IR (KBr): ν = 3359, 3062, 3023, 2971, 2930, 1742, 1703,
˜
1453, 1381, 1132, 1053, 700 cm–1. CIMS: m/z (%) = 350.2 (1)
[M+H+], 225.8 (4), 167.0 (12) [Ph2CH+], 88.9 (10), 59.0 (100). HR-
CIMS: calcd. for C22H24NO3 350.1756, found 350.1767 [M+H+].
1
crude product (10.87 g, 70.7% yield) as a colourless oil. H NMR
(200 MHz, CDCl3): δ = 1.36 (d, J = 6.2 Hz, 3 H), 4.19 (dq, J =
4.3, 6.2 Hz, 1 H), 4.31 (d, J = 4.3 Hz, 1 H), 5.37 (br. s, 1 H) ppm.
13C NMR (50 MHz, CDCl3): δ = 19.9, 52.6, 67.3, 172.1 ppm. IR
(1ЈR,3S,4S)-4-(Cyclopropylcarbonyl)-3-(1-hydroxyethyl)azetidin-2-
one (23e): NBS[62] (0.99 g, 5.6 mmol, 1.1 equiv.) was added to a
mixture of 13e (1.8 g, 5.1 mmol, 1 equiv.) and Br2 (1.3 mL of a
solution of 0.1 mL Br2 in 10 mL of CH2Cl2 ; 0.25 mmol,
0.05 equiv.) in CH2Cl2 (70 mL) and water (45 mL). The mixture
was stirred at room temperature overnight, under irradiation from
the fume hood lamp (discoloration occurred). After dilution with
CH2Cl2 and extraction with NaHSO3 (5%), the organic phase was
(film): ν = 3438, 2983, 2936, 1721, 1408, 1381, 1283, 1125,
˜
1026 cm–1. EIMS: m/z (%) = 139.8 (24) [CH3CH(OH)CH281Br+],
137.9 (24) [CH3CH(OH)CH279Br+], 121.7(7), 119.9 (5), 45.0 (100).
(2R,3R)-2,3-Epoxybutyrate Mixed Salt 2b: A solution of NaOH
(4.19 g, 103 mmol, 2 equiv.) in water (16 mL) was added dropwise,
over 60 min, to a cold (–12 °C) solution of bromothreonine (9.42 g,
52 mmol, 1 equiv.) in water (7 mL). The solution was allowed to
warm to 20 °C, stirred for 18 h, diluted with acetone (50 mL) and
concentrated under reduced pressure. The residue was dried under
high vacuum to furnish 2b (mixed salt) as a white solid (11.56 g,
1
concentrated to furnish crude 22e (1.87 g, 100% yield). H NMR
(200 MHz, CDCl3): δ = 0.7–1.0 (m, 4 H), 1.28 (d, J = 6.2 Hz, 3
H), 1.85 (tt, J = 4.8, 7.7 Hz, 1 H), 2.95 (dd, J = 2.5, 5.1 Hz, 1 H),
4.22 (dq, J = 5.1, 5.8 Hz, 1 H), 4.62 (d, J = 2.5 Hz, 1 H), 5.63 (br.
s, 1 H), 7.1–7.8 (m, 10 H) ppm. 13C NMR (50 MHz, CDCl3): δ =
11.7, 12.4, 17.8, 21.8, 60.5, 60.7, 64.5, 87.6, 126.4–132.3, 141.7,
166.7, 209.7 ppm. Crude 22e (1.87 g, 5.1 mmol) was dissolved in
acetone/water (1:1, 70 mL) containing PTSA (0.94 g, 1.9 mmol,
0.95 equiv.). The mixture was stirred in the dark at room tempera-
ture overnight. Concentration under vacuum gave a brown solid
(2.77 g) containing PTSA, benzophenone and 23e. Benzophenone
1
97.9% yield). H NMR (200 MHz, D2O): δ = 1.28 (d, J = 5.2 Hz,
3 H), 3.34 (dq, J = 5.1, 5.2 Hz, 1 H), 3.53 (d, J = 5.1 Hz, 1 H)
ppm. 13C NMR (50 MHz, D2O): δ = 15.2, 55.8, 58.5, 177.8 ppm.
IR (KBr): ν = 3083, 3012, 2965, 2933, 1611, 1441, 1O36, 911,
˜
671 cm–1.
(2R,3R)-N-Benzhydryl-N-(2-cyclopropyl-2-oxoethyl)-2,3-epoxy-
butyramide (8e): Oxalyl chloride (1.43 mL, 16 mmol, 1.4 equiv.) was
3764
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 3755–3766

Synthesis of a Novel Azetidin-2-one Carbapenem Precursor
FULL PAPER
[4]
[5]
[6]
T. Nakatsuka, H. Iwata, R. Tanaka, S. Imajo, M. Ishiguro, J.
Chem. Soc., Chem. Commun. 1991, 662–664.
N. Sayo, T. Saito, Y. Okeda, H. Nagashima, H. Kumobayashi,
U. S. Patent US4981992, 1991.
R. Noyori, T. Ikeda, T. Ohkuma, M. Widhalm, M. Kitamura,
H. Takaya, S. Akutagawa, N. Sayo, T. Asito, T. Taketomi, H.
Kumobayashi, J. Am. Chem. Soc. 1989, 111, 9134–9135.
S. Hanessian, A. Bedeschi, C. Battistini, N. Mongelli, J. Am.
Chem. Soc. 1985, 107, 1438–1439.
could be recovered by precipitation from cold water (90–100 %
yield). Pure N-deprotected azetidinone 23e was obtained by column
chromatography on RP-18 (CH3CN/H2O, 1:3) as a yellow foam
(0.77 g, 82% yield, Table S1). RF = 0.76 (RP-18; CH3CN/H2O, 1:3;
1
visualization with phosphomolybdic acid). H NMR (200 MHz,
CDCl3): δ = 1.04 (dt, J = 2.9, 7.9 Hz, 2 H), 1.14 (dt, J = 2.9, 4.5 Hz,
2 H), 1.41 (d, J = 6.3 Hz, 3 H), 2.17 (tt, J = 4.5, 7.9 Hz, 1 H), 3.22
(ddd, J = 1.6, 2.7, 4.4 Hz, 1 H), 4.33 (dq, J = 4.4, 6.3 Hz, 1 H),
4.45 (d, J = 2.7 Hz, 1 H), 6.56 (br. s, 1 H) ppm. 13C NMR (75 MHz,
CDCl3): δ = 12.0, 12.2, 17.2, 21.9, 57.3, 63.7, 65.0, 167.2,
[7]
[8]
D. R. Kronenthal, C. Y. Han, M. K. Taylor, J. Org. Chem.
1982, 47, 2765–2768.
207.6 ppm. IR (film): ν = 3309, 3012, 2971, 2930, 1750, 1701, 1397,
[9]
˜
P. J. Kocienski, Protecting Groups, 3rd ed., Georg Thieme,
Stuttgart, New York, 2004.
B.-L. Deng, M. Demillequand, M. Laurent, R. Touillaux, M.
Belmans, L. Kemps, M. Cérésiat, J. Marchand-Brynaert, Tetra-
hedron 2000, 56, 3209–3217.
1127, 994 cm–1. CIMS: m/z (%) = 184.3 (5) [M+H+], 148.8 (23),
71.6 (34), 60.6 (96), 58.7 (100). HR-CIMS: calcd. for C9H14NO3
184.0974, found 184.0978 [M+H+].
[10]
(1ЈR,3S,4S)-4-(Cyclopropylcarbonyloxy)-3-(1-hydroxyethyl)azetidin-
2-one (24e): A mixture of 23e (0.26 g, 1.44 mmol, 1 equiv.) and
mCPBA[63] (0.49 g, 2.8 mmol, 2 equiv.) in CH2Cl2 (10 mL) was
stirred at room temperature overnight. Concentration and column
chromatography on silica gel (CH2Cl2/ethyl acetate, 3:2) afforded
24e (0.25 g, 89 % yield) as a white solid (Table S1). 1H NMR
(300 MHz, [D6]acetone): δ = 0.9–1.0 (m, 4 H), 1.27 (d, J = 6.3 Hz,
3 H), 1.68 (m, 1 H), 3.15 (dd, J = 1.5, 5.7 Hz, 1 H), 4.09 (dq, J =
5.7, 6.3 Hz, 1 H), 5.90 (s, 1 H), 7.98 (br. s, 1 H) ppm. 13C NMR
(75 MHz, [D6]acetone): δ = 9.0, 9.1, 13.2, 22.1, 64.3, 66.0, 76.5,
[11]
[12]
[13]
[14]
M. Laurent, M. Belmans, L. Kemps, M. Cérésiat, J. Marchand-
Brynaert, Synthesis 2003, 570–576.
M. Laurent, M. Cérésiat, J. Marchand-Brynaert, J. Org. Chem.
2004, 69, 3194–3197.
H. Yanagisawa, A. Ando, M. Shiozaki, T. Hiraoka, Tetrahe-
dron Lett. 1983, 24, 1037–1040.
M. Shiozaki, H. Maruyama, N. Ishida, Heterocycles 1984, 22,
1725–1726.
[15]
[16]
Y. Petit, C. Sanner, M. Larchevêque, Synthesis 1988, 538–540.
F. Grosjean, M. Huché, M. Larchevêque, J. J. Legendre, Y. Pe-
tit, Tetrahedron 1994, 50, 9325–9334.
Y. Shimohigashi, M. Waki, N. Izumiya, Bull. Chem. Soc. Jpn.
1979, 52, 949–950.
M. Demillequand, M. Cérésiat, M. Belmans, L. Kemps, J.
Marchand-Brynaert, J. Chromatogr. A 1999, 832, 259–264.
M. Laurent, J. Marchand-Brynaert, Synthesis 2000, 667–672.
M. Kugelman, D. Gala, R. S. Jaret, P. L. Nyce, A. T. McPhail,
Synlett 1990, 431–432.
166.8, 175.0 ppm. IR (KBr):ν = 3311, 3017, 2973, 2933, 1769, 1732,
˜
[17]
[18]
1378, 1160, 1034, 862 cm–1. CIMS: m/z (%) = 199.9 (25) [M+H+],
182.0 (47), 155.9 (100), 113.9 (44), 86.6 (69), 70 (65). HR-CIMS:
calcd. for C9H14NO4 200.0923, found 200.0920 [M+H+].
[19]
[20]
(1ЈR,3S,4S)-3-[1-(tert-Butyldimethylsilyloxy)ethyl]-4-(cyclopropyl-
carbonyloxy)azetidin-2-one (1e): TBDMSCl (0.076 g, 0.5 mmol,
4 equiv.) and imidazole (0.09 g, 1.3 mmol, 12 equiv.) were added to
a solution of 24e (0.022 g, 0.11 mmol, 1 equiv.) in dry DMF
(3 mL). The mixture was stirred at 20 °C for 2 d. After dilution
with ethyl acetate, the solution was washed with brine (3 times),
dried with MgSO4 and concentrated. The residue was dried under
high vacuum to give 1e (0.033 g, 95% yield) as a white solid
(Table S1). 1H NMR (300 MHz, CDCl3): δ = 0.05 (s, 3 H), 0.07 (s,
3 H), 0.85 (s, 9 H), 0.94 (ddd, J = 3.6, 4.5, 7.8 Hz, 2 H), 1.03 (dt,
J = 3.6, 4.5 Hz, 2 H), 1.25 (d, J = 6.3 Hz, 3 H), 1.63 (tt, J = 4.5,
7.8 Hz, 1 H), 3.19 (dd, J = 1.2, 3.6 Hz, 1 H), 4.22 (dq, J = 3.6,
6.3 Hz, 1 H), 5.83 (s, 1 H), 6.48 (br. s, 1 H) ppm. 13C NMR
(75 MHz, CDCl3): δ = –4.22, 9.14, 9.24, 12.8, 18.0, 22.4, 25.7, 63.9,
[21]
[22]
[23]
[24]
[25]
[26]
D. H. R. Barton, C.-Y. Chern, C. Tachdjian, Heterocycles 1994,
37, 793–805.
J. Hartung, R. Kneuer, M. Schwarz, I. Svoboda, H. Fuess, Eur.
J. Org. Chem. 1999, 97–106.
B. Tinant, J.-P. Declercq, M. Cérésiat, J. Marchand-Brynaert,
Z. Kristallogr. – New Cryst. Struct. 2001, 216, 51–52.
B. Tinant, J.-P. Declercq, M. Cérésiat, J. Marchand-Brynaert,
Z. Kristallogr. – New Cryst. Struct. 2001, 216, 53–54.
B. Tinant, M. Laurent, M. Cérésiat, J. Marchand-Brynaert, Z.
Kristallogr. – New Cryst. Struct. 2003, 218, 145–147.
M. Kugelman, D. Gala, R. S. Jaret, P. L. Nyce, A. T. McPhail,
Synlett 1990, 431–432.
J. E. Baldwin, J. Chem. Soc., Chem. Commun. 1976, 734–736.
J. E. Baldwin, L. I. Kruse, J. Chem. Soc., Chem. Commun.
1977, 233–235.
J. E. Baldwin, M. L. Lusch, Tetrahedron 1982, 38, 2939–2947.
S. J. Gorzynski, Synthesis 1984, 629–656.
[27]
[28]
65.0, 75.0, 166.4, 175.0 ppm. IR (KBr): ν = 3111, 2954, 2930, 2857,
˜
1779, 1735, 1392, 1161, 1076, 837 cm–1. CIMS: m/z (%) = 428.4 (8)
[M+SiMe2tBu+], 298.1 (17) [M – CH3+], 269.9 (20), 256.0 (30)
[M – C4H9+], 228.1 (43) [M – OCOC3H5+], 159.0 (100). HR-CIMS:
[29]
[30]
[31]
calcd. for C11H22NSiO2 228.1420, found 228.1416 [M
–
F. Benedetti, F. Berti, S. Fabrissin, T. Gianferrara, J. Org.
Chem. 1994, 59, 1518–1524.
OCOC3H5+].
[32]
[33]
P. Crotti, V. Di Bussolo, L. Favero, F. Macchia, M. Pineschi,
E. Napolitano, Tetrahedron 1999, 55, 5853–5866.
P. Crotti, F. Badalassi, V. Di Bussolo, L. Favero, M. Pineschi,
Tetrahedron 2001, 57, 8559–8572.
Supporting Information (see footnote on the first page of this arti-
cle): Physical constants and analysis.
[34]
[35]
D.-F. Huang, L. Huang, Tetrahedron 1990, 46, 3135–3142.
F. Wu, M. G. Buhendwa, D. F. Weaver, J. Org. Chem. 2004, 69,
9307–9309.
J. N. Be Miller, R. E. Wing, C. Y. Meyers, J. Org. Chem. 1968,
33, 4292–4294.
H. Hashimoto, M. Kawa, Y. Saito, T. Date, S. Horito, J. Yoshi-
mura, Tetrahedron Lett. 1987, 28, 3505–3508.
M. S. Anson, J. G. Montana, Synlett 1994, 219–220.
S. R. Baker, A. F. Parsons, M. Wilson, Tetrahedron Lett. 1998,
39, 331–332.
R. Riess, M. Schön, S. Laschat, V. Jäger, Eur. J. Org. Chem.
1998, 473–479.
Acknowledgments
This work was supported by the F.N.R.S. (Fonds National de la
Recherche Scientifique, Belgium) and Tessenderlo Chemie (Tessen-
derlo, Belgium). J. M.-B. is a senior research associate of F.N.R.S.
[36]
[37]
[38]
[39]
[1] O. A. Mascaretti, Bacteria versus Antibacterial Agents: an Inte-
grated Approach, ASM Press, Washington, DC, 2003.
[2] A. H. Berks, Tetrahedron 1996, 52, 331–375.
[3] T. Osashi, K. Kan, N. Ueyama, I. Sada, A. Miyama, K. Wat-
anabe, U. S. Patent US 4791198, 1988.
[40]
Eur. J. Org. Chem. 2006, 3755–3766
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3765

M. Laurent, M. Cérésiat, J. Marchand-Brynaert
FULL PAPER
[41] C. M. Gasparki, A. Ghosh, M. J. Miller, J. Org. Chem. 1992,
57, 3546–3550.
[42] J. C. Bussolari, K. Beers, P. Lalan, W. V. Murray, D. Gauthier,
P. Mc Donnell, Chem. Lett. 1998, 787–788.
[43] C. Yu, L. Hu, J. Org. Chem. 2002, 67, 219–223.
[55] D. W. Heaton, I. M. Cunnigham, Eur. Patent, EP276993, 1988.
[56] K. Fujimoto, H. Maruyama, Japan Patent Appl. JP02304072,
1990.
[57] S. Chackalamannil, N. Fett, M. Kirkup, A. Afonso, A. K.
Ganguly, J. Org. Chem. 1988, 53, 450–452.
[44] A. P. Patwardhan, Z. Lu, V. R. Pulgam, W. D. Wulff, Org. Lett. [58] M. Shiozaki, N. Ishida, T. Hiraoka, H. Maruyama, Tetrahe-
2005, 7, 2201–2204.
[45] G.-J. Ten Brink, I. W. C. E. Arends, R. A. Sheldon, Chem. Rev.
2004, 104, 4105–4123.
[46] M. Renz, B. Meunier, Eur. J. Org. Chem. 1999, 737–750.
[47] C. M. Crudden, A. C. Chen, L. A. Calhoun, Angew. Chem. Int.
Ed. 2000, 39, 2851–2855.
[48] O. Reiser, Chemtracts: Org. Chem. 2001, 14, 94–99.
[49] A. Arrieta, B. Lecea, F. P. Cossio, C. Palomo, J. Org. Chem.
1988, 53, 3784–3791.
[50] M. Altamura, M. Ricci, Synth. Commun. 1988, 18, 2129–2133. [63] N. N. Schwartz, J. H. Blumbergs, J. Org. Chem. 1964, 29, 1976–
[51] S.-I. Murahashi, Y. Oda, T. Naota, Tetrahedron Lett. 1992, 33,
7557–7560.
[52] M. Shiozaki, Synthesis 1990, 691–693.
[53] B. Alcaide, M. F. Aly, M. A. Sierra, J. Org. Chem. 1996, 61,
8819–8825.
dron 1984, 40, 1795–1802.
[59] M. J. Calverley, Tetrahedron 1987, 43, 4609–4619.
[60] Y. Tsuchiya, T. Nomoto, M. Hayashi, Y. Iwasawa, H. Masaki,
M. Ohkubo, Y. Sakuma, Y. Nagata, T. Satoh, T. Kamei, Eur.
Patent Appl. EP448078, 1991.
[61] F. H. Amondo-Neizer, R. A. Shaw, D. O. Skolvin, B. C. Smith,
J. Chem. Soc. 1965, 2997–2999.
[62] S. C. Virgil, in Encycl. Reagents Org. Synth. (Ed.: L. A. Pa-
quette), John Wiley, Chichester, 1995, vol. 1, p. 768.
1978.
[64] B. Tinant, M. Laurent, J. Marchand-Brynaert, Z. Kristallogr. –
New Cryst. Struct. 2006, 221, 68–70.
Received: March 17, 2006
Published Online: June 14, 2006
[54] M. G. Rosenberg, U. Haslinger, U. H. Brinker, J. Org. Chem.
2002, 67, 450456.
3766
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 3755–3766