Exploring the chemistry of penicillin as a β-lactamase-dependent prodrug

Article abstract of DOI:10.1039/b614758e

The penam nucleus can be modified to behave as a β-lactamase-dependent 'prodrug' by incorporation of a vinyl ester side chain at the 6-position. Enzyme-catalysed hydrolysis of the β-lactam ring uncovers the thiazolidine-ring nitrogen as a nucleophile that

Full text of DOI:10.1039/b614758e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Exploring the chemistry of penicillin as a b-lactamase-dependent prodrug†
Carol C. Ruddle and Timothy P. Smyth*
Received 10th October 2006, Accepted 8th November 2006
First published as an Advance Article on the web 22nd November 2006
DOI: 10.1039/b614758e
The penam nucleus can be modified to behave as a b-lactamase-dependent ‘prodrug’ by incorporation
of a vinyl ester side chain at the 6-position. Enzyme-catalysed hydrolysis of the b-lactam ring uncovers
the thiazolidine-ring nitrogen as a nucleophile that drives a rapid intramolecular displacement on the
side chain. Attachment of 7-hydroxy-4-methylcoumarin as the releasable group of this side chain
generated a penicillin structure that can function as a fluorescence-based reporter substance/diagnostic
for the presence of low levels of b-lactamase enzyme in solution. Mechanistic details of the reaction
pattern are documented and the scope and limitations of exploiting the structural modification are
discussed.
Introduction
b-Lactams remain the frontline antibiotics for combating infec-
tions caused by Gram-negative bacteria. Their efficacy, however,
is being compromised increasingly by the dissemination of b-
lactamase enzymes throughout this bacterial population.1 While
many serine-based b-lactamases are susceptible to inhibition by
compounds such as clavulanic acid and sulbactam, the emergence
of metallo-b-lactamases poses a new challenge, as these are not
susceptible to such inhibitors, and the newer generation of these
enzymes act on a broad variety of b-lactam structures—penicillins,
cephalosporins and carbapenems. b-Lactamase-dependent pro-
drugs provide a potential strategy to counter the spread of b-
lactamases by exploiting their presence, rather than blocking
their action.2 Both the cephalosporin and carbapenem nucleus
incorporate an inherent reactivity pattern that allows these to
be readily configured as b-lactamase-dependent prodrugs. Thus,
elimination of the 3ꢀ-substituent in cephalosporins (Scheme 1a),
and of the 2ꢀ-substituent in carbapenems, is triggered by scission
of the b-lactam; this reactivity has been explored mainly in
the development of cephalosporin–quinolone conjugates.3,4 De-
spite the accumulation of detailed biological activity data, the
cephalosporin–quinolone conjugates have not been launched as
therapeutic agents.3f Nonetheless, the strategy of developing b-
Scheme 1 b-Lactams showing a latent prodrug reactivity pattern: (a) a
lactamase-dependent prodrugs remains valid and, in this context,
we have explored the chemistry of the penam nucleus with the
aim of incorporating a latent ‘prodrug’ reactivity that is triggered
by scission of the b-lactam ring. A generic modification that
achieves this is attachment of an unsaturated linker, with an
electron-deficient centre bearing a moderately nucleofugic group,
at the 6-position of penicillin. Structures 1 and 2 are prototypic
examples: enzyme-catalysed scission of the b-lactam ring brings
two potential nucleophiles into play—the thiazolidine-ring sulfur
and nitrogen atoms—resulting in a rapid intramolecular reaction
typical cephalosporin, (b) generic penicillin structures.
with loss of the nucleofugic group (Scheme 1b).5,6 The efficacy of
the displacement is due to the syn juxtaposition of the nucleophile
and electron-deficient centre on the unsaturated linker6 of the
ring-opened structure, leading to formation of a five-membered
ring: this type of molecular construct gives rise to reactions of
very high effective molarity.7 It proved difficult to prepare S-
aminosulfenimino structures (Scheme 1, 1) with leaving groups
other than a small set of arylsulfonamides.8 As a result we
turned our attention to penicillin structures bearing a vinylester
side chain. Penicillin structures bearing this side chain, and with
the thiazolidine ring sulfur in the sulfone oxidation state, have
previously been found to act as b-lactamase inhibitors, whereas
the corresponding sulfides are not inhibitors.9a In the context of
our work, the vinyl ester side chain was seen as an alternative to
the S-aminosulfenimine moiety and one offering better scope
Department of Chemical and Environmental Sciences, University of Lim-
erick, National Technological Park, County Limerick, Ireland. E-mail:
timothy.smyth@ul.ie
† Electronic supplementary information (ESI) available: 1H NMR spectral
sequences for reaction of 7ꢀ and 20ꢀ with b-lactamase enzyme; Gaussian
energy values for addition of methoxide to the a-face of an ethylidene
b-lactam unit. See DOI: 10.1039/b614758e
160 | Org. Biomol. Chem., 2007, 5, 160–168
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
for synthetic variation of the leaving group. In a preliminary
communication6 we reported on the preparation and reaction
of a variety of compounds of structure type 2 bearing distinct
leaving groups; in this paper we expand on the chemistry of these
penicillins and on the scope and limitations of their capability to
act as b-lactamase-dependent prodrugs.
6 the reaction stopped after ring opening (6a was isolated and
characterised);6 it would appear that the size of t-butyl group
impedes the intramolecular displacement. The b-lactam ring of
the ethylidene structure 7 was remarkably stable and remained
totally unchanged under the above reaction conditions; ester
exchange of the benzhydryl group was the only process observed
after 3 h. This stability was also observed for the corresponding
sodium salt 7ꢀ in aqueous buffer (see below) and indicates that the
exocyclic methyl group, juxtaposed syn to the lactam carbonyl,
provides a significant barrier to nucleophilic attack. Indeed, the
tetrahedral intermediate formed by addition of methoxide to the
a-face of an ethylidene b-lactam unit was found to be 2.1 kcal
mol−1 less stable (optimised structures, 6–311G level, Gaussian 03)
than the corresponding methylidene intermediate, when compared
with the respective starting materials.10 Such a difference in
activation energies would lead to a reduction in reaction rate of
over two orders of magnitude. (The ethylidene structural feature
is present in asparenomycins11—naturally occurring carbapenem
structures—and may contribute to their stability.)
Results and discussion
An efficient route to structure type 2 is via reaction of oxo-
penicillin derivative3 with a Wittig reagent (Scheme 2) as originally
developed by Buynak.9 The benzhydryl esters 4–7 were prepared
using commercially available Wittig reagents and, where required,
the corresponding sodium salts were obtained by removal of the
benzhydryl group.6 Preparation of 8ꢀ was achieved by removal
of the allyl group of 5 followed by DCC mediated coupling of
the resulting free acid with 7-hydroxy-4-methylcoumarin (below
referred to as coumarin) and, finally, removal of the benzhydryl
group. This structure is a potential reporter molecule/diagnostic
for the presence of b-lactamase enzyme as 8ꢀ is non-fluorescent,
whereas coumarin itself is highly fluorescent.
The reaction pattern shown in Scheme 3a was initially validated
for the benzhydryl esters 4 and 5 in a reaction system of CH3OH–
0.2 M NEt3 for 40 min at r.t. (nominally 18 ◦C); the ring-fused
c-lactam 10 was formed exclusively and was fully characterised.6
This isomer, with the unsaturated bridgehead position, was found
to be 3.74 kcal mol−1 more stable (optimised structures, 6–
311G level, Gaussian 03) than 9; the isomerisation is driven,
presumably, by initial deprotonation at the bridgehead carbon
of 9 in the basic reaction medium. With the t-butyl derivative
The reactivity patterns of the salts 4ꢀ, 7ꢀ and 8ꢀ in aqueous buffer
(D2O, pH 7.2, 25 ◦C) both in the absence and presence of b-
lactamase enzyme were examined and the results are presented
below. In the absence of enzyme 4ꢀ had a half-life of 5 days.
The reaction products were identical to those of the enzyme-
catalysed hydrolysis (see following), which confirms that the b-
lactam ring is hydrolysed more readily that the aliphatic ester
group. The time course of the reaction of 4ꢀ in the presence
1
of b-lactamase enzyme was followed by H NMR spectroscopy
(see Fig. 1). The spectral sequence shows the initial formation of
an intermediate and conversion of this to a single final product
Scheme 2 Synthetic route: (i) ROC(O)CR PPh3, −55 ◦C; (ii) (a) AlCl3, −84 ◦C, (b) NaHCO3, freeze dry; (iii) (a) 5, Pd(PPh3)4, sodium p-toluene
ꢀ
=
sulfinate, THF–H2O, (b) 7-hydroxy-4-methylcoumarin, DCC.
Scheme 3 Reaction pattern of: (a) benzhydryl esters in CH3OH–0.2 M NEt3, (b) sodium salt 4ꢀ in D2O buffer (pH 7.2), 25 ◦C in the presence of
b-lactamase enzyme.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 160–168 | 161
©

View Article Online
Fig. 1 1H NMR spectra of (a) 4ꢀ in D2O buffer (0.02 M phosphate, pH 7.2, 25 ◦C) and (b) 7.5 min, (c) 38 min, (d) 50 min following addition of
b-lactamase enzyme in a citrate buffer.
with the co-production of methanol. The key indicator that the
intermediate is the ring-opened structure 4ꢀa, is the large upfield
shift in the a-CH3 resonance6 (Fig. 1b,c); there is also a small
upfield shift in the methyl ester resonance and a distinct change in
the H-5 and H-8 resonances. The conversion of 4ꢀa to 9ꢀ with the
concomitant product of methanol is clearly seen in the sequence of
spectra (b) → (c) → (d) (Fig. 1). For the end product (Fig. 1d), the
downfield shift of the a-methyl resonance (compared with that of
the ring-opened structure 4ꢀa) is notable and is consistent with the
re-formation of a ring-fused bicyclic structure, while the relatively
short reaction time for the whole transformation precludes simple
hydrolysis of the methyl ester group of 4ꢀa. The 13C-NMR and
high-resolution mass spectra of the product isolated after workup
are consistent with structure 9ꢀ.
The ethylidene salt 7ꢀ showed remarkable stability in aqueous
buffer and was unchanged after 14 days (this parallels the earlier
observation on the stability of b-lactam ring of ester 7 in CH3OH–
0.2M NEt3). Nonetheless, b-lactamase-catalysed hydrolysis of 7ꢀ
was effective and this led cleanly to the formation of 9ꢀꢀ and the
co-production of ethanol. (The formation of 9ꢀꢀ confirms the Z
stereochemistry of the side chain of 7ꢀ.) During the course of the
reaction build up, the corresponding ring-opened structure was
not observed.10 It is interesting to note that the vinyl methyl group,
which exerts such a dramatic effect on nucleophilic attack at the
lactam carbonyl, does not impede the intramolecular nucleophilic
displacement at the side chain carbonyl.
b-Lactamase-catalysed hydrolysis of 8ꢀ led cleanly to the for-
mation of 9ꢀ and the co-release of coumarin; the corresponding
ring-opened structure was not observed during the course of the
reaction. In the absence of enzyme 8ꢀ was much more labile than
the aliphatic derivative 4ꢀ, giving a half-life of 90 min in the
buffer system. This lability was associated with direct hydrolysis
of the aromatic ester side chain rather than with hydrolysis of
the b-lactam ring. The 1H NMR spectra of the end products
of hydrolysis of 8ꢀ in the absence and presence of b-lactamase
enzyme are shown in Fig. 2: in the presence of enzyme the
products were coumarin and c-lactam 9ꢀ, while in the absence
of enzyme the products were coumarin and b-lactam 11 (the
high-resolution mass spectra of the product isolated after workup
Unlike the case of the benzhydryl esters 4 and 5, the c-lactam
isomer obtained here was 9ꢀ; the aqueous reaction medium was
not sufficiently basic to deprotonate the bridgehead carbon atom
and, in addition, the carboxylate side chain should impede the
deprotonation.
162 | Org. Biomol. Chem., 2007, 5, 160–168
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Fig. 2 1H NMR spectra of the hydrolysis products of 8ꢀ (a) 50 min after addition of b-lactamase enzyme, and (b) after 12 h in D2O buffer (0.02 M
phosphate, pH 7.2, 25 ◦C). The peaks labelled C correspond to free 7-hydroxy-4-methylcoumarin in both spectra.
Table 1 Hydrolytic stability and relative efficiency as b-lactamase sub-
strates of 4ꢀ, 7ꢀ and 8ꢀ
Relative rate of turnover Half-life in 0.02 M phosphate
Structure
by b-lactamase enzymea buffer, pH 7.2, 25 ◦
C
4ꢀ
7ꢀ
8ꢀ
1
0.5
20
7 daysc
ꢁ14 daysd
1.5 he
Benzylpenicillin 1260b
—
a Penicillinase Type 1 from B. cereus. b The specific activity with benzylpeni-
cillin was 1872 lmol min−1 mg−1 protein. c Hydrolysis of the b-lactam ring.
d No hydrolysis occurred within 14 days. e Hydrolysis of the aryl ester side
chain.
is consistent with structure 11 (see the Experimental section)).
A summary of some of the key data on 4ꢀ, 7ꢀ and 8ꢀ is given in
Table 1.
The release of coumarin from 8ꢀ was also monitored by
fluorescence spectroscopy. The time course of the fluorescence
emission of a 1.4 mM solution of 8ꢀ (phosphate buffer, pH 7.2),
when treated with various amounts of b-lactamase enzyme, is
shown in Fig. 3. It is clear that 8ꢀ can act as a reporter molecule
for the presence of quite low amounts (∼6 units)12 of b-lactamase
enzyme in solution. When 8ꢀ was treated with a solution containing
b-lactamase-producing E. coli cells13 the rate of fluorescence
emission was not faster than that associated with background
hydrolysis of 8ꢀ itself, whereas, when treated with cell lysate
the rate of fluorescence emission was considerably faster. These
observations indicate that 8ꢀ cannot efficiently permeate the outer
membrane of E. coli.
Fig. 3 Increase in fluorescence emission as a function of time observed
for a 1.4 mM solution of 8ꢀ in the presence of (a) 0, (b) 6, (c) 30, and
(d) 60 units of b-lactamase enzyme (penicillinase type 1).
buffer was a drawback as it restricted the type of releasable prodrug
group that could be usefully incorporated and evaluated (e.g.
triclosan). It was envisaged that an aryl sulfonate side chain would
avoid this limitation and that such a structure might be obtainable
via a Horner–Wittig–Emmons reaction of the required sulfophos-
phonate with 3 (Scheme 4). Whereas this method was successful
in producing the alkyl sulfonate derivative 16 (26% yield) it
did not yield the aryl sulfonate 17. The problem was the lack
of reactivity of phenylmethanesulfonate anion as a nucleophile
in general, and specifically toward diphenylchlorophosphate14 to
produce 15. Treatment of 13 with n-BuLi followed by CH3I failed
Although aryl esters are more readily hydrolysed than aliphatic
esters, the lability of the ester side chain of 8ꢀ, compared to that of
4ꢀ, was unexpected. The relatively short half-life of 8ꢀ in aqueous
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 160–168 | 163
©

View Article Online
to produce any methylated product (a similar process with 12
rapidly produced methylethane sulfonate). The breakdown of 13
in the presence of n-BuLi to generate sulfene15 and phenoxide was
not the reason for this lack of nucleophilic behaviour as treatment
of 13 with n-BuLi followed by quenching with D2O lead to the
recovery of a substantial amount of deuterated phenylmethane
sulfonate. As a similar problem is likely to be associated with the
anion of 15, this approach was not pursued further. A phosphonate
side chain was considered as a final alternative. Treatment of 3
with the anion derived from tetraethylbisphosphonate produced
the penicillin derivative 18; as this was obtained in only a 1% yield
(after purification) no further evaluation was undertaken.
as the nucleophile providing this assistance, and not the imine
nitrogen; the coproduct(s) were not identified. Treatment of the
desacetoxy cephalosporin salt 21ꢀ with b-lactamase enzyme led
quantitatively to the release of methanol and the formation of the
c-lactam 22 (Scheme 6); the ring-opened structure was observed
here as it was with 4ꢀa.10 With a half-life of 55 min for the
displacement the enamine nitrogen of 21ꢀa is more nucleophilic,
toward the side chain carbonyl, than the dihydrothiazine sulfur,
but is less nucleophilic by about a factor of 5, that the amine
nitrogen in the ring-opened penicillin 4ꢀa.
Scheme 4 (i) (a) n-BuLi, (b) Cl(O)P(OPh)2, −78 ◦C; (ii) (a), n-BuLi,
(b) 3, −78 ◦C.
Scheme 6 c-Lactam formation from desacetoxy cephalosporin structure
21.
c-Lactam salt 23, obtained from 10 by removal of the benzhydryl
group, was screened for antibiotic and b-lactamase-inhibitory
activity. While no antibiotic activity was observed, 23 was found to
show weak inhibition (50% at 100 lM) of the metallo-b-lactamase
CphA from Aeromonas hydrophilia.16
The release of methanol from the cephalosporin salts 19ꢀ and 20ꢀ
was also studied and the key findings are summarised in Scheme 5.
Cephalosporin 19ꢀ was as stable as the corresponding penicillin
4ꢀ with respect to background hydrolysis and also behaved as a
b-lactamase substrate. Loss of methanol from the ring-cleaved
structure 19ꢀa was considerably faster than that observed in
background hydrolysis, thus indicating anchimeric assistance. The
finding that loss of methanol from the ring-cleaved sulfoxide 20ꢀa
was slower than from 19ꢀa, identifies the dihydrothiazine sulfur
Conclusions
Engineering the penicillin nucleus to act as a viable b-lactamase-
dependent prodrug presents a distinct challenge compared with
that of achieving the same capacity in cephalosporins. Ideally, the
intact structure must behave as an antibiotic in its own right and, in
addition, the ‘prodrug’ component must behave as an independent
antibiotic when released following scission of the b-lactam ring.
The 7-amino- and 3ꢀ-substituents of cephalosporins makes it easier
to achieve this: the 7-amino substituent (not-releasable) can be
used to obtain normal b-lactam antibiotic activity, while the 3ꢀ-
substituent can be chosen to act as an independent antibiotic when
released. In the case of a penicillin bearing either a vinyl ester or
S-aminosulfenimine side chain (Scheme 1; 1 and 2, respectively),
this group must allow for normal b-lactam antibiotic activity of
the intact structure, and it must also carry the releasable moiety
that can function as an independent antibiotic. In all cases the
releasable antibiotic component must have biological activity and
clearance profiles that are compatible with those of a b-lactam;
this is a concern in the case of the cephalosporin–quinolone
conjugates.5c In our work we have identified the chemical basis
Scheme 5 Rates of methanol release from cephalosporin structures 19ꢀ
and 20ꢀ.
164 | Org. Biomol. Chem., 2007, 5, 160–168
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
of modifying the penam unit so as to incorporate a latent prodrug
reactivity that is triggered by the action of a b-lactamase enzyme.
One functional implementation of this is the coumarin derivative
8ꢀ which can act as a fluoresence-based reporter molecule for the
presence of low levels of b-lactamase enzyme in solution.
and the resulting aqueous layer was lyophilised to leave 7ꢀ as a
yellow solid (67.2 mg, 0.209 mmol, 85%) dH (270 MHz, D2O) 1.30
(t, J = 7.2 Hz, 3H, CH2CH3) 1.48 (s, 3H, a-CH3), 1.55 (s, 3H,
b-CH3), 2.11 (s, 3H, CH3), 4.18–4.31 (m, 2H, CH2CH3) 4.33 (s,
1H, H-3), 5.96 (s, 1H, H-5).
Sodium-6-(Z)-(methylumbelliferyl)carbonylmethyl-
Experimental
idenepenicillanate (8ꢀ)
Benzhydryl ester 86 (135 mg, 0.232 mmol) was treated as described
above. The free acid of 8ꢀ (73 mg, 0.175 mmol, 76%), recovered as a
pale yellow solid, was dissolved in ethyl acetate and extracted with
aqueous sodium hydrogen carbonate (13.25 mg, 0.157 mmol) and
the resulting aqueous layer was lyophilised to leave 8ꢀ as a yellow
solid. dH (270 MHz, D2O) 1.54 (s, 3H, a-CH3), 1.61 (s, 3H, b-CH3),
2.51 (s, 3H, CH3Ar), 4.46 (s, 1H, H-3), 6.17 (s, 1H, H-5), 6.42 (s,
Benzhydryl-6-(Z)-[1-ethoxycarbonylethylidene]penicillanate (7)
Benzhydryl-6-oxopenicillinate 36 (1.04 g, 2.73 mmol) was dis-
solved in dichloromethane (13 mL). The solution was cooled
to −55 ◦C under nitrogen. To this cooled solution (1-ethoxy-
carbonylethyledene)triphenylphosphorane (931.4 mg, 2.57 mmol)
in dichloromethane (30 mL) was added dropwise over 30 min.
Stirring was continued for a further 10 min. Analysis by TLC
(50 : 50, ethyl acetate–hexane) indicated disappearance of the
3 and the formation of a new product of higher Rf value. The
solution was allowed warm to room temperature and was washed
with water (20 mL), the organic layer was separated, dried and
concentrated under reduced pressure to leave crude 7. Purification
by chromatography (silica gel, hexane–ethyl acetate 60 : 40) yielded
7 as a waxy, yellow oil (737 mg, 1.63 mmol, 63%). ESI-HRMS
for C26H27NO5S: [M + H]+1 calcd: 466.1688, found: 466.1709. dH
(270 MHz, CDCl3) 1.27 (s, 3H, a-CH3), 1.33 (t, J = 7.2 Hz,
3H, CH2CH3) 1.56 (s, 3H, b-CH3), 2.20 (s, 3H, CH3), 4.25 (q,
J = 7.2 Hz, 2H, CH2CH3) 4.64 (s, 1H, H-3), 5.95 (s, 1H, H-5),
6.93 (s, 1H, CH(Ar)2), 7.22–7.42 (m, 10H, ArH); dC (67.93 MHz,
=
=
1H, C(O)CH C, side chain), 6.66 (s, 1H CH C, coumarin), 7.27
(dd, J = 1.86, 8.91 Hz, 1H, ArH, coumarin), 7.33 (d, J = 1.86 Hz,
ArH, coumarin), 7.89 (d, J = 8.91 Hz, 1H, ArH, coumarin).
Diphenylphosphorylmethanesulfonate (14)
Methylmethane sulfonate (800 lL, 9.44 mmol) was dissolved in
dry THF (24 mL) and the resultant solution was purged with
nitrogen for 5 min. The solution was then cooled to −78 ◦C
and 2 M n-BuLi in pentane (5.2 mL, 10.38 mmol) was added.
Diphenylchlorophosphate (1076 lL, 5.2 mmol) was added after
stirring the butyllithium solution at −78 ◦C for 15 min. The
resultant solution was maintained at −78 ◦C for 30 min and
then at −50 ◦C for 60 min. Ammonium chloride (555.44 mg,
10.58 mmol) in H2O (6 mL) was added and the mixture was
warmed to room temperature. The mixture was diluted with water
(50 mL) and extracted with dichloromethane (2 × 60 mL). The
combined organic extracts were washed with water (3 × 40 mL),
dried over MgSO4, filtered and concentrated to yield a white solid
(1.82 g). The crude material was recrystallised from diethyl ether–
dichloromethane to give the 14 as a white solid (1.04 g, 3.03 mmol,
58%). Mp 82–83.5 ◦C; found C, 48.87; H, 4.23%. C14H15O6PS
requires C, 49.12; H, 4.42%; dH (270 MHz, CDCl3) 4.04 (s, 3H,
CH3), 4.04 (d, J = 17.32 Hz, 2H, –CH2–P), 7.20–7.34 (m, 10H,
ArH); dC (75.47 MHz, CDCl3) 46.94 (d, J = 144.28 Hz, CH2-
P, inverted in DEPT135), 57.68 (CH3O), 120.61, 120.54, 126.03,
126.01, 130.06, 130.04, (C, Ar), 149.56, 149.69 (C, Ar).
=
CDCl3) 14.13, 14.45, 25.59 (CH3CH2O, CH3CH , a-CH3), 33.67
(b-CH3), 61.77, 63.41, 69.89, 70.34 (C-2, C-3, C-5, CH3CH2O),
78.31 (CH(Ar)2), 127.03, 127.52, 128.16, 128.34, 128.55, 128.62 (C,
Ar), 139.17, 139.23 (C, Ar), 149.05 (C-6), 165.10, 167.02, 167.82
=
(3 × C O).
Sodium-6-(Z)-[1-ethoxycarbonylethylidene]penicillanate (7ꢀ)
7 (200 mg, 0.442 mmol) was dissolved in dichloromethane
(16 mL) and was cooled under nitrogen to −84 ◦C. A solution
of aluminium trichloride (240 mg, 1.80 mmol) in nitromethane
(1.5 mL) was added in one portion to the cooled penicillin solution
at which point the solution became intensely yellow. After stirring
for 20 min, ethyl acetate (75 mL) and 5% sodium hydrogen
carbonate (75 ml) were added successively whilst maintaining the
temperature at −84 ◦C. The resulting slushy mixture was allowed
to reach room temperature, the two layers were separated and the
aqueous layer was filtered through celite. The filtrate was extracted
with ethyl acetate (20 ml). The aqueous extract was collected and
ethyl acetate (30 mL) was layered on top, the pH was adjusted to
2.2, and extracted with the ethylacetate. The aqueous layer was
extracted with a second portion of ethyl acetate (30 mL). The
organic extracts were combined, dried and concentrated under
reduced pressure to leave the free acid of 7 as a pale yellow solid
(98 mg, 0.305 mmol, 69%). ESI-HRMS for C13H17NO5S: [M–H]−1
calcd: 298.0749, found: 298.0744; dH (300 MHz, CDCl3) 1.35 (t,
J = 6.9 Hz, 3H, CH2CH3) 1.58 (s, 3H, a-CH3), 1.62 (s, 3H, b-CH3),
2.21 (s, 3H, CH3), 4.27 (m, J = 6.9 Hz, 2H, CH2CH3) 4.57 (s, 1H,
H-3), 5.93 (s, 1H, H-5), 8.11 (br s, 1H, OH). The free acid of 7
(83 mg, 0.258 mmol) was dissolved in ethyl acetate and extracted
with aqueous sodium hydrogen carbonate (20.6 mg, 0.245 mmol)
Benzhydryl-6-(Z)-methoxysulfonatemethylidenepenicillanate (16)
Diphenylphosphorylmethanesulfonate (14) (1.51 g, 2.57 mmol)
was dissolved in THF (8 ml) and cooled under nitrogen atmo-
◦
sphere to −78 C. 2 M n-BuLi (1.28 ml, 2.57 mmol) was added
◦
and the mixture was stirred for 15 minutes at −78 C. 3 (1.04 g,
2.73 mmol) was added to the cold mixture and was allowed to
stir at this temperature for a further 10 min. Analysis by TLC
(50 : 50, ethyl acetate–hexane) indicated disappearance of the oxo-
compound and the formation of a new product of higher Rf value.
The solution was allowed warm to room temperature and was
washed with water (20 mL), the organic layer was separated, dried
and concentrated under reduced pressure to leave crude 16 as
a brown oil. Purification by chromatography (silica gel, hexane–
ethyl acetate, 70 : 30) yielded 16 (339.5 mg, 0.657 mmol, 26%).
◦
Mp 54–56 C; found C, 58.42; H, 4.92; N, 2.94%. C23H23NO6S2
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 160–168 | 165
©

View Article Online
requires C, 58.34; H, 4.90; N, 2.96%; mmax (KBr)/cm−1 1779, 1745;
dH (270 MHz, CDCl3) 1.26 (s, 3H, a-CH3), 1.57 (s, 3H, b-CH3), 3.92
(s, 3H, –OCH3), 4.47 (s, 1H, H-3), 6.04 (s, 1H, H-5), 6.68 (s, 1H, CH
side chain), 6.95 (s, 1H, CH(Ar)2), 7.32–7.36 (m, 10H, ArH); dC
(67.93 MHz, CDCl3) 25.26 (a-CH3), 33.73 (b-CH3), 57.47 (CH3O),
64.36, 68.46, 70.82 (C-2, C-3, C5), 78.68 (CH(Ar)2), 118.63 (side
4.57 (d, J = 12.62 Hz, 1H, H-3ꢀA), 4.64 (d, J = 12.62 Hz, 1H,
H-3ꢀB), 5.24 (d, J = 1.48 Hz, 1H, H-4), 5.74 (d, J = 1.48 Hz,
1H, H-6), 6.33 (d, J = 0.99 Hz, 1H, H-2), 6.48 (d, J = 0.99 Hz,
1H, C(O)CH), 6.91 (s, 1H, CH(Ar)2), 7.25–7.38 (m, 10H, ArH); dC
(67.93 MHz, CDCl3) 20.75 (–C(O)CH3), 51.40, 52.64, 53.67, 65.76
(C-4, CH3O, C-6, C-3ꢀ), 79.28 (CH(Ar)2), 114.99 119.15, 124.26 (C-
=
=
chain C(O)C ), 127.04, 127.53, 128.31, 128.50, 128.63, 128.69 (C,
3, C-2, C(O)CH ), 126.85, 127.19, 128.40, 128.54, 128.76, 128.84,
Ar), 138.79, 138.89 (C, Ar), 156.08 (C-6), 163.98, 163.99, 166.19
(C Ar), 138.95, 138.99 (C Ar), 155.03 (C-7), 158.10, 164.21, 166.46,
3
=
=
(3 × C O).
170.52 (4 × C O). See below for characterisation data on the D
isomer.
Benzhydryl-6-(Z)-diethoxyphosphonatemethylidenepenicillanate
(18)
Sulfoxide formation
Tetraethylmethylenediphosphonate (425.7 lL, 1.71 mmol) was
dissolved in THF (8 mL) and cooled under nitrogen atmosphere
to −78 ◦C. 2 M n-BuLi (0.856 mL, 1.71 mmol) was added and
the mixture was stirred for 15 minutes at −78 ◦C. 3 (0.98 g,
2.56 mmol) was added to the cold mixture and was allowed to
stir at this temperature for a further 10 minutes. Analysis by
TLC (50 : 50, ethyl acetate–hexane) indicated disappearance of
the oxo-compound. The solution was allowed warm to room
temperature and was washed with water (20 mL), the organic layer
was separated, dried and concentrated under reduced pressure
to leave a brown oil. Purification by chromatography (silica gel,
hexane–ethyl acetate, 70 : 30) gave a mixture of products. This
purification step was repeated to yield a pure sample 18 (∼10 mg,
∼1%). ESI-HRMS for C26H30NO6PS: [M + Na]+1 calcd: 538.1429,
found: 538.1422; dH (270 MHz, CDCl3) 1.26 (s, 3H, a-CH3), 1.35
(t, J = 5.44 Hz, 3H, –(OCH2CH3)A), 1.36 (t, J = 5.44 Hz,
3H, –(OCH2CH3)B), 1.56 (s, 3H, b-CH3), 4.08–4.21 (m, 4H –
(OCH2CH3)A overlapping with –(OCH2CH3)B), 4.64 (s, 1H, H-3),
5.98 (t, J = 1.24 Hz, 1H, H-5), 6.24 (dd, J = 1.24, 13.36 Hz, 1H,
CH sidechain), 6.95 (s, 1H, CH(Ar)2), 7.23–7.44 (m, 10H, ArH).
The D2/D3 mixture, (485.7 mg, 0.98 mmol) was dissolved in
dichloromethane (29 ml) and the solution was cooled to 0 ◦C.
m-Chloroperoxybenzoic acid, 86% (256.6 mg, 1.49 mmol) was
dissolved in dichloromethane (4 mL) and was added dropwise
to the cold cephalosporin solution. The solution was stirred at
0 ◦C for 40 min and reaction progress was monitored by TLC. A
cold solution of 10% sodium sulfite (7.1 mL) and 0.1 M sodium
hydrogen carbonate (7.1 mL) was added and the biphasic mixture
was stirred at 0 ◦C for 10 min. The layers were separated. The
organic layer was washed with 0.1 M sodium hydrogen carbonate
(3 × 7 mL) and 20% NaCl solution (10 mL), dried over magnesium
sulfate and concentrated under vacuum to yield the benzhydryl
sulfoxide ester 20—a yellow solid (349.2 mg, 0.685 mmol, 70%
yield) as a mixture of a/b-sulfoxide isomers. For characterisation
purposes (see below) a small sample of each isomer was obtained
by chromatography (silica gel, ethyl acetate–hexane, 80 : 20); for
the subsequent reduction process the isomer mixture of 20 was
used. b-Isomer (20) (less polar material):17 found C, 61.27; H, 4.59;
N, 2.60%. C26H23NO8S requires C, 61.29; H, 4.55; N, 2.75%; dH
(270 MHz, CDCl3) 2.08 (s, 3H, –C(O)CH3), 3.55 (d, J = 16.49 Hz,
1H, H-2A), 3.84 (d, J = 16.49 Hz, 1H, H-2B), 3.87 (s, 3H, CH3O),
4.99 (d, J = 14.65 Hz, 1H, H-3ꢀA), 5.16 (d, J = 14.65 Hz, 1H,
H-3ꢀB), 5.26 (d, J = 1.47 Hz, 1H, H-6), 6.62 (d, J = 1.47 Hz, 1H,
C(O)CH), 6.97 (s, 1H, CH(Ar)2), 7.25–7.50 (m, 10H, ArH); dC
Benzhydryl-7-[(Z)-(methoxycarbonyl)methylidene)]-
cephalosporanate-1-sulfoxide (20)
Benzhydryl-7-oxocephalosporanate, prepared as described for
(67.93 MHz, CDCl3) 20.66 (–C(O)CH3), 51.15, 52.97, 62.00 (C-2,
ꢀ
3,6 (1.027 g, 2.35 mmol) was dissolved in dichloromethane
CH3O, C-3 , C-6), 80.10 (CH(Ar)2), 119.82 (C(O)CH ), 126.86,
=
◦
(15 mL). The solution was cooled to −84 C under nitrogen. To
127.24, 127.36, 128.31, 128.34, 128.58, 128.66, (C-3, C-4,C Ar),
this cooled solution methyl(triphenylphosphoranylidene)acetate
(740 mg, 2.21 mmol) in dichloromethane (30 mL) was added
dropwise over 30 min. Stirring was continued for a further 10 min.
Analysis by TLC (50 : 50, ethyl acetate–hexane) indicated disap-
pearance of the oxo-compound and the formation of a new prod-
uct of higher Rf value. The solution was allowed warm to room
temperature and was washed with water (20 mL), the organic layer
was separated, dried and concentrated under reduced pressure to
leave crude as a brown oil. Purification by chromatography (silica
gel, hexane–ethyl acetate, 70 : 30) yielded the benzhydryl ester Wit-
tig product benzhydryl-7-[(Z)-(methoxycarbonyl)methylidene)]-
cephalosporanate (19)—a yellow–green glassy gel (485.7 mg,
0.984 mmol, 45% yield)—as a mixture of D2 and D3 isomers
(approx. 70 : 30): For characterisation purposes (see below) a small
sample of each isomer was obtained by chromatography (silica
gel, hexane–ethyl acetate, 70 : 30); for the subsequent sulfoxide
formation and reduction—required to obtain a larger sample of
the D3 isomer—the isomer mixture of 19 was used. D2 isomer: dH
(270 MHz, CDCl3) 1.96 (s, 3H, –C(O)CH3), 3.82 (s, 3H, CH3O),
131.48, 138.84 (C Ar), 148.41 (C-7), 157.12, 159.33, 163.40, 170.23
=
(4 × C O). a-Isomer (20) (more polar material): found C, 60.90;
H, 4.53; N, 2.55%. C26H23NO8S requires C, 61.29; H, 4.55; N,
2.75%; dH (270 MHz, CDCl3) 2.02 (s, 3H, –C(O)CH3), 3.30 (d,
J = 19.42 Hz, 1H, H-2A), 3.84 (d, J = 19.42 Hz, 1H, H-2B), 3.85
(s, 3H, CH3O), 4.69 (d, J = 13.92 Hz, 1H, H-3ꢀA), 5.21 (d, J =
13.92 Hz, 1H, H-3ꢀB), 5.18 (d, J = 1.47 Hz, 1H, H-6), 6.63 (d, J =
1.47 Hz, 1H, C(O)CH), 7.00 (s, 1H, CH(Ar)2), 7.26–7.50 (m, 10H,
ArH).
Benzhydryl-7-[(Z)-(methoxycarbonyl)methylidene)]-
cephalosporanate (19, D3 isomer) by sulfoxide reduction
The a/b-isomer mixture of 20 (182.4 mg, 0.358 mmol) was dis-
solved in DMF (8 ml) and cooled to −40 ◦C. Phosphorus trichlo-
ride (63 lL, 0.72 mmol) was added. The mixture was allowed
to warm from −40 to −10 ◦C while stirring under nitrogen. The
reaction mixture was added to cold water (21 mL) and the product
19 (D3 isomer) precipitated out immediately. The precipitate,
166 | Org. Biomol. Chem., 2007, 5, 160–168
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
=
a yellow solid, was collected by vacuum filtration, washed with
water and dried over P2O5 for 3 days, (142.2 mg, 0.288 mmol, 80%).
Found C, 63.15; H, 4.61; N, 2.67%. C26H23NO7S requires C, 63.28;
H, 4.70; N, 2.84%; mmax (KBr)/cm−1 1781, 1732; dH (300 MHz,
CDCl3) 2.02 (s, 3H, –C(O)CH3), 3.39 (d, J = 18.60 Hz, 1H, H-2A),
3.62 (d, J = 18.60 Hz, 1H, H-2B), 3.84 (s, 3H, CH3O), 4.76 (d, J =
13.51 Hz, 1H, H-3ꢀA), 4.98 (d, J = 13.51 Hz, 1H, H-3ꢀB), 5.49 (d,
J = 1.20 Hz, 1H, H-6), 6.48 (d, J = 1.20 Hz, 1H, C(O)CH), 6.99
(s, 1H, CH(Ar)2), 7.22–7.48 (m, 10H, ArH).
79.20 (CH(Ar)2), 116.36 (C(O)CH ), 123.80, 127.14, 127.55,
128.02, 128.06, 128.47, 128.56 (C-3, C Ar), 133.23, 139.52, 139.64
=
(C-4, C Ar), 153.59 (C-7), 157.75, 161.32, 164.21 (3 × C O).
This ester was treated with aluminium chloride as described
above for 7ꢀ to remove the benzhydryl group and give the free acid
of 21ꢀ (40 mg, 0.149 mmol, 69%): ESI-HRMS for C11H11NO5S:
[M–H]−1 calcd: 268.0280, found: 268.0292; dH (270 MHz, CDCl3)
2.26 (s, 3H, CH3), 3.29 (d, J = 18.56 Hz, 1H, H-2A), 3.62 (d, J =
18.56 Hz, 1H, H-2B), 3.84 (s, 3H, OCH3), 4.83 (br s, 1H, OH), 5.51
(d, J = 0.62 Hz, 1H, H-6), 6.46 (d, J = 0.99 Hz, 1H, C(O)CH).
The free acid (30.3 mg, 0.113 mmol) was dissolved in ethyl acetate
and extracted with aqueous sodium hydrogen carbonate (8 mg,
0.096 mmol) and the resulting aqueous layer was lyophilised to
leave 21ꢀ as a bright yellow solid (22 mg, 0.076 mmol, 79%); mmax
(KBr)/cm−1 1746 (lactam), 1704; dH (270 MHz, D2O) 1.95 (s, 3H,
CH3), 3.27 (d, J = 18.31 Hz, 1H, H-2A), 3.70 (d, J = 18.06 Hz, 1H,
H-2B), 3.81 (s, 3H, OCH3), 5.61 (s, 1H, H-6), 6.52 (s, 1H, C(O)CH).
Sodium-7-(Z)-methoxycarbonylmethylidenecephalosporanate (19ꢀ)
Removal of the benzhydryl ester of 19 (D3 isomer), (86 mg,
0.174 mmol), was effected as described above for 7ꢀ above. The salt
19ꢀ was obtained as a pale yellow solid (40.1 mg, 0.115 mmol, 66%).
Free acid: ESI-HRMS for C13H13NO7S: [M–H]−1 calcd: 326.0033,
found: 326.0319; mmax (KBr)/cm−1 3440, 1784, 1773, 1723; dH
(270 MHz, CDCl3) 2.12 (s, 3H, –C(O)CH3), 3.45 (d, J = 18.80 Hz,
1H, H-2A), 3.67 (d, J = 18.80 Hz, 1H, H-2B), 3.85 (s, 3H, CH3O),
4.96 (d, J = 13.85 Hz, 1H, H-3ꢀA), 5.15 (d, J = 13.85 Hz, 1H, H-3ꢀB),
5.52 (s, 1H, H-6), 6.50 (s, 1H, C(O)CH); dC (67.93 MHz, CDCl3),
Fluorescence measurement
Coumarin salt 8ꢀ (0.124 mg, 2.8 × 10−4 mmol) in 0.02 M phosphate
◦
dC 20.80(–C(O)CH3), 27.86, 52.67, 57.85, 63.27 (C-2, C-6, CH3O,
buffer (200 lL), pH = 7.2 at 25 C was combined with varying
ꢀ
=
C-3 ), 80.10 (CH(Ar)2), 117.99 (C(O)CH ), 126.24, 126.90, 151.90
amounts (0, 6, 30, 60 units) of penicillinase type 1. The fluorescence
emission at 448 nm (excitation wavelength 360 nm) was recorded
as a function of time on a TECAN, GENios Pro instrument.
=
(C-3, C-4, C-7), 158.09, 163.16, 163.91, 171.20 (4 × C O). Sodium
salt 19ꢀ: dH (270 MHz, D2O) 2.11 (s, 3H, –C(O)CH3), 3.43 (d, J =
18.3 Hz, 1H, H-2A), 3.73 (d, J = 18.3 Hz, 1H, H-2B), 3.84 (s, 3H,
CH3O), 4.74 (d, J = 12.6 Hz, 1H, H-3ꢀA), 4.89 (d, J = 12.6 Hz,
1H, H-3ꢀB), 5.65 (s, 1H, H-6), 6.58 (s, 1H, C(O)CH).
Kinetic analysis
1H NMR spectra were recorded in D2O buffers (0.02 M phosphate,
pH 7.2) at 25 ◦C. Penicillinase type 1 from B. cereus (Sigma) was
used for enzyme-catalysed hydrolysis.
Sodium-7-(Z)-methoxycarbonylmethylidenecephalosporanate
sulfoxide (b-isomer) (20ꢀ)
Removal of the benzhydryl ester of 20 (b-isomer), (137.4 mg,
0.270 mmol), was effected as described above for 7ꢀ above. For
the free acid: dH (270 MHz, CDCl3) 2.13 (s, 3H, –C(O)CH3), 3.64
(d, J = 16.58 Hz, 1H, H-2A), 3.88 (s, 3H, CH3O), 3.98 (d, J =
16.58 Hz, 1H, H-2B), 5.03 (d, J = 14.60 Hz, 1H, H-3ꢀA), 5.21 (d,
J = 14.35 Hz, 1H, H-3ꢀB), 5.39 (d, J = 1.24 Hz, 1H, H-6), 6.62 (d,
J = 1.48 Hz, 1H, C(O)CH). The salt 20ꢀ was obtained as a pale
yellow solid (42.2 mg, 0.115 mmol, 43%). dH (270 MHz, D2O) 2.12
(s, 3H, –C(O)CH3), 3.82 (d, J = 16.34 Hz, 1H, H-2A), 3.83 (s, 3H,
CH3O), 4.23 (d, J = 16.34 Hz, 1H, H-2B), 5.61 (d, J = 1.22 Hz,
1H, H-6), 6.74 (d, J = 1.22 Hz, 1H, C(O)CH).
2,2-Dimethyl-5-oxo-2,3,5,7a-tetrahydro-pyrrolo[2,1-b]thiazole-3,
7-dicarboxylic acid (9ꢀ)
See ref. 6 for full characterisation (1H, 13C NMR spectra and ESI-
HRMS).
2,2,6-Trimethyl-5-oxo-2,3,5,7a-tetrahydro-pyrrolo[2,1-b]-
thiazole-3,7-dicarboxylic acid (9ꢀꢀ)
Following completion of the kinetic analysis run of the b-
lactamase-catalysed hydrolysis of 7ꢀ: dH (270 MHz, D2O) 1.56
(s, 3H, a-CH3), 1.57 (s, 3H, b-CH3), 2.02 (d, J = 1.5 Hz, 3H,
CH3), 4.16 (s, 1H, H-3), 5.90 (s, 1H, H-8). After lyophilisation:
ESI-HRMS for C11H13NO5S: [M–H]−1 calcd: 270.0436, found:
270.0432.
Sodium-7-(Z)-methoxycarbonylmethylidenedesacetoxy-
cephalosporanate (21ꢀ)
Benzhydryl-7-(Z)-methoxycarboylmethylidenedesacetoxycepha-
losporanate (21) was prepared, starting from commercially avail-
able7-aminodesacetoxycephalosporanic acid, bythe methodology
given above for 19. The benzhydryl ester (367 mg, 50%) was
obtained as a mixture of D2 and D3 isomers (37 : 63) which
were separated by chromatography. D3 isomer: mp 58–60 ◦C;
ESI-HRMS for C24H22NO5S: [M + H]+1 calcd: 436.1219, found:
436.1235; mmax (KBr)/cm−1 1781 (lactam), 1732; dH (270 MHz,
CDCl3) 2.12 (s, 3H, CH3), 3.23 (d, J = 18.06 Hz, 1H, H-2A),
3.51 (d, J = 18.06 Hz, 1H, H-2B), 3.83 (s, 3H, OCH3), 5.48 (d,
J = 1.73 Hz, 1H, H-6), 6.44 (d, J = 1.24 Hz, 1H, C(O)CH)
6.96 (s, 1H, CH(Ar)2), 7.21–7.51 (m, 10H, ArH); dC (67.93 MHz,
CDCl3), 20.22 (CH3), 31.69 (C-2), 52.49 (OCH3), 57.96 (C-6),
6-(Z)-Carboxymethylidenepenicillanic acid (11)
Following completion (12 h) of the buffer-catalysed hydrolysis of
8ꢀ: dH (270 MHz, D2O) 1.48 (s, 3H, a-CH3), 1.54 (s, 3H, b-CH3),
4.28 (s, 1H, H-3), 5.92 (s, 1H, H-5), 6.31 (s, 1H, C(O)CH C,
sidechain). After lyophilisation: ESI-HRMS for C10H11NO5S:
[M–H]−1 calcd: 256.0280, found: 256.0285.
=
3-Methyl-6-oxo-6,8a-dihydro-2H-pyrrolo[2,1-b][1,3]thiazine-4,
8-dicarboxylic acid (22)
Following completion of the kinetic analysis run of the b-
lactamase-catalysed hydrolysis of 21ꢀ: dH (270 MHz, D2O) 1.95
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 160–168 | 167
©

View Article Online
(s, 3H, CH3), 3.29 (d, J = 17.81 Hz, 1H, H-2A), 3.73 (d, J =
17.81 Hz, 1H, H-2B), 5.31 (s, 1H, H-9), 6.60 (s, 1H, H-7). Distilled
water (20 mL) was added, the solution was acidified to pH 2.2 and
extracted with ethyl acetate (2 × 20 mL). The organic layer was
dried and concentrated under reduced pressure to leave the free
acid of 22 as a yellow solid: ESI-HRMS for C10H9NO5S: [M–H]−1
calcd 254.0123 found 254.0130; dH (270 MHz, (CD3)2CO) 2.06 (s,
3H, CH3), 2.96 (br s, 1H, OH), 3.40 (d, J = 17.81 Hz, 1H, H-2A),
3.84 (d, J = 17.81 Hz, 1H, H-2B), 5.48 (d, J = 0.87 Hz, 1H, H-9),
6.92 (d, J = 1.24 Hz, 1H, H-7).
3 (a) H. A. Albrecht, G. Beskid, K.-K. Chan, J. G. Christenson, R.
Cleeland, K. H. Deitcher, N. H. Georgopapadakou, D. D. Keith,
D. L. Pruess, J. Sepinwall, A. C. Specian, R. L. Then, M. Weigele,
K. F. West and R. Yang, J. Med. Chem., 1990, 33, 77–86; (b) H. A.
Albrecht, G. Beskid, J. G. Christenson, J. W. Durkin, V. Fallat, N. H.
Georgopapadakou, D. D. Keith, F. M. Konzelmann, E. R. Lipschitz,
D. H. McGarry, J.-A. Siebelist, C. C. Wei, M. Weigele and R. Yang,
J. Med. Chem., 1991, 34, 669–675; (c) H. A. Albrecht, G. Beskid, J. G.
Christenson, N. H. Georgopapaakou, D. D. Keith, F. M. Konzelmann,
D. L. Pruess, P. L. Rossman and C. C. Wei, J. Med. Chem., 1991, 34,
2857–2864; (d) H. A. Albrecht, G. Beskid, J. G. Christenson, K. H.
Dietcher, N. H. Georgopapadakou, D. D. Keith, F. M. Konzelmann,
D. L. Pruess and C. C. Wei, J. Med. Chem., 1994, 37, 400–407; (e) T. P.
Demuth, R. E. White, R. A. Tietjen, R. J. Storrin, J. R. Skuster, J. A.
Andersen, C. C. McOsker, R. Freedman and F. J. Rourke, J. Antibiot.,
1991, 44, 200–209; (f) P. M. Hershberger and T. P. Demuth, Adv. Exp.
Med. Biol., 1998, 456, 239–267 and references therein.
2,2-dimethyl-5-oxo-2,3,5,6-tetrahydropyrrolo[2,1-b]thiazole-3,
7-dicarboxylic acid 7-methyl ester (23)
4 An interesting exploitation of this elimination reaction pattern has been
reported in the development of a carbapenem antibiotic: H. Rosen, R.
Hajdu, L. Silver, H. Kropp, K. Dorso, J. Kohler, J. G. Sundelof, J. Huber,
G. G. Hammond, J. J. Jackson, C. J. Gill, R. Thompson, B. A. Pelak,
J. H. Epstein-Toney, G. Lankas, R. R. Wilkening, K. J. Wildonger, T. A.
Blizzard, F. P. DiNinno, R. W. Ratcliffe, J. V. Heck, J. W. Kozarich and
M. L. Hammond, Science, 1999, 283, 703–706.
5 (a) T. P. Smyth, M. E. O’Donnell, M. J. O’Connor and J. O. St, Ledger,
J. Org. Chem., 1998, 63, 7600–7618; (b) T. P. Smyth, M. J. O’Connor and
M. E. O’Donnell, J. Org. Chem., 1999, 64, 3132–3138; (c) T. P. Smyth,
M. E. O’Donnell, M. J. O’Connor and J. O. St, Ledger, Tetrahedron,
2000, 56, 5699–5707.
The benzhydryl ester 10 was treated as described above for
7ꢀ. The free acid of 23 was obtained as a yellow glassy solid
(49 mg, 0.181 mmol, 66%); mp 51–52 ◦C (dec); ESI-HRMS
for C11H13NO5S: [M–H]−1 calcd 270.0436 found 270.0439; mmax
(KBr)/cm−1 3438, 1733, 1687; dH (270 MHz, CDCl3) 1.65 (s, 3H,
a-CH3), 1.76 (s, 3H, b-CH3), 3.55 (d, J = 23.26 Hz, 1H, H-6A),
3.65 (d, J = 23.26 Hz, 1H, H-6B), 3.75 (s, 3H, OCH3), 4.42 (s, 1H,
H-3), 6.00 (br s, 1H, OH); dC (67.93 MHz, CDCl3), 25.30 (a-CH3),
32.99 (b-CH3), 40.27 (CH2), 51.52 (CH3O), 62.41, 65.00 (C-2, C-
=
3), 97.21 (C-7), 156.27 (C-8), 164.20, 169.32, 173.31 (3 × C O).
6 C. C. Ruddle and T. P. Smyth, Chem. Commun., 2004, 2332–2333.
7 (a) A. J. Kirby, in Advances in Physical Organic Chemistry, vol. 17,
ch. 3, 1980; (b) L. Mandolini, in Advances in Physical Organic
Chemistry, vol. 22, ch. 1, 1986.
The free acid (31 mg, 0.114 mmol) was dissolved in ethyl acetate
and extracted with aqueous sodium hydrogen carbonate (9.1 mg,
0.108 mmol) and the resulting aqueous layer was lyophilised to
leave as a dark green solid (27.4 mg, 0.093 mmol, 82%): dH
(270 MHz, D2O) 1.59 (s, 3H, a-CH3), 1.72 (s, 3H, b-CH3), 3.64
(s, 2H, CH2), 3.76 (s, 3H, OCH3), 4.29 (s, 1H, H-3).
8 J. W. Grant and T. P. Smyth, J. Org. Chem., 2004, 69, 7965–7970.
9 (a) J. D. Buynak, B. Geng, B. Bachmann and L. Hua, Bioorg. Med.
Chem. Lett., 1995, 5, 1513–1518; (b) J. D. Buynak, A. S. Rao and S.
Nidamarthy, Tetrahedron Lett., 1998, 39, 4945–4946; (c) J. D. Buynak,
WO Pat. 99/33837, 1999; (d) J. D. Buynak, K. Wu, B. Bachmann, D.
Khasnis, L. Hua, H. K. Nguyen and C. L. Carver, J. Med. Chem.,
1995, 38, 1022–1034; (e) J. D. Buynak, V. R. Doppalapudi and G.
Adam, Bioorg. Med. Chem. Lett., 2000, 10, 853–857.
Acknowledgements
10 See data in the ESI†.
We are grateful to R. O’Dwyer of the University of Limerick
for preparation of the b-lactamase-producing E. coli culture
and for assistance with the fluorescence measurements, and to
C. J. Schofield, B. Lienard and W. Sobey of the University of
Oxford, UK, and to L. Horsfall and J.-M. Fre`re of the Universite´
de Lie`ge, Belgium for the antibiotic and b-lactamase inhibition
assays. We are grateful to D. Rai of the CSCB, University College
Dublin for the HRMS, and to IRCSET for financial support
(C.C.R.).
11 (a) N. Tsuji, K. Nagashima, M. Kobayashi, J. Shoji, T. Kato, Y. Terui,
H. Nakai and M. Shiro, J. Antibiot., 1982, 35, 24–31; (b) Y. Kimura, K.
Motokawa, H. Nagata, Y. Kameda, S. Matsuura, M. Mayama and T.
Yoshida, J. Antibiot., 1982, 35, 32–38; (c) K. Murakami, M. Doi and
T. Yoshida, J. Antibiot., 1982, 35, 38–45.
12 One unit hydrolyzes 1.0 lmole of benzylpenicillin per min at pH 7.0 at
25 ◦C.
13 The E. coli strain Top10 (Invitrogen) containing the plasmid pIGXH,
expressing an ampicillin resistant gene, was used. This material
was prepared by R. O’Dwyer in the Department of Chemical and
Environmental Sciences, University of Limerick.
14 See ref. 10 in: B. E. Maryanoff, M. J. Constanzo, S. O. Nortey, M. N.
Greco, R. P. Shank, J. J. Schupsky, M. P. Ortegon and J. L. Vaught,
J. Med. Chem., 1998, 41, 1315–1343.
15 (a) T. Ritter, K. Stanek, I. Larrosa and E. M. Carreira, E. M., Org.
Lett., 2004, 6, 1513–1514; (b) A. F. Barrero, J. E. Oltra, M. Alvarez and
A. Rosales, J. Org. Chem., 2002, 67, 5461–5469.
References
1 T. A. Murphy, L. E. Catto, S. E. Halford, A. T. Hadfield, W. Minor, T. R.
Walsh and J. Spencer, J. Mol. Biol., 2006, 357, 890–903 and references
therein.
2 (a) C. H. O’Callaghan, R. B. Sykes and S. E. Staniforth, Antimicrob.
Agents Chemother., 1976, 245–248; (b) D. Greenwood and F. O’Grady,
Antimicrob. Agents Chemother., 1976, 249–252; (c) S. Mobashery and
M. Johnson, J. Biol. Chem., 1986, 261, 7879–7887; (d) S. Mobashery
and M. Johnson, Biochemistry, 1987, 26, 5878–5884; (e) T. P. Smyth,
M. E. O’Donnell, M. J. O’Connor and J. O. St Ledger, Tetrahedron,
2000, 56, 5699–5707 and references therein.
16 The screening assays were carried out at the Chemistry Re-
search Laboratory, University of Oxford, UK, and at the Centre
d’Inge´nierie des Proteines, Universite´ de Lie`ge, Belgium, under the
supervision of Prof. C. J. Schofield, and Prof. J.-M. Fre`re, res-
pectively.
17 For details of sulfoxide isomer assignment see: J. B. Doherty, et al.,
J. Med. Chem., 1990, 33, 2513–2521.
168 | Org. Biomol. Chem., 2007, 5, 160–168
This journal is
The Royal Society of Chemistry 2007
©