Antibiotic resistance: Mono- and dinuclear zinc complexes as metallo-β-lactamase mimics

Article abstract of DOI:10.1002/chem.200600629

Biomimetic systems containing one or two zinc(II) ions supported by phenolate ligands were developed as functional mimics of metallo-β- lactamase. These complexes were shown to catalytically hydrolyze β-lactam- substrates, such as oxacillin and penicillin G. The dinuclear zinc complex 1, which has a coordinated water molecule, exhibits high β-lactamase activity, whereas the dinuclear zinc complex 2, which has no water molecules, but labile chloride ligands, shows a much lower activity. The high β-lactamase ac tivity of complex 1 can be ascribed to the presence of a zinc-bound water molecule that is activated by being hydrogen bonded to acetate substituents. The kinetics of the hydrolysis of oxacillin by complex 1 and the effect of pH on the reaction rates are reported in detail. In addition, the kinetic parameters obtained for the synthetic ana logues are compared with those of the natural metallo-β-lactamase from Bacillus cereus (Bell). To understand the role of the second metal ion in hydrolysis, the syntheses and catalytic activities of two mononuclear complexes (3 and 4) that include coordinated water molecules are described. Interestingly, the mononuclear zinc complexes 3 and 4 also exhibit high activity, supporting the assumption that the second zinc ion is not crucial for the β-lactamase activity.

Full text of DOI:10.1002/chem.200600629

FULL PAPER
DOI: 10.1002/chem.200600629
Antibiotic Resistance: Mono- and Dinuclear Zinc Complexes as Metallo-b-
Lactamase Mimics
A. Tamilselvi, Munirathinam Nethaji, and G. Mugesh*[a]
Dedicated to Professor K. C. Nicolaou on the occasion of his 60th birthday
Abstract: Biomimetic systems contain-
ing one or two zinc(II) ions supported
by phenolate ligands were developed
as functional mimics of metallo-b-lacta-
mase. These complexes were shown to
catalytically hydrolyze b-lactam sub-
strates, such as oxacillin and penicil-
lin G. The dinuclear zinc complex 1,
which has a coordinated water mole-
cule, exhibits high b-lactamase activity,
whereas the dinuclear zinc complex 2,
which has no water molecules, but
labile chloride ligands, shows a much
lower activity. The high b-lactamase ac-
tivity of complex 1 can be ascribed to
the presence of a zinc-bound water
molecule that is activated by being hy-
drogen bonded to acetate substituents.
The kinetics of the hydrolysis of oxacil-
lin by complex 1 and the effect of pH
on the reaction rates are reported in
detail. In addition, the kinetic parame-
ters obtained for the synthetic ana-
logues are compared with those of the
natural metallo-b-lactamase from Ba-
cillus cereus (BcII). To understand the
role of the second metal ion in hydroly-
sis, the syntheses and catalytic activities
of two mononuclear complexes (3 and
4) that include coordinated water mol-
ecules are described. Interestingly, the
mononuclear zinc complexes 3 and 4
also exhibit high activity, supporting
the assumption that the second zinc ion
is not crucial for the b-lactamase activi-
ty.
Keywords: antibiotics · bioinorgan-
ic chemistry · enzyme models · pen-
icillin · zinc
Introduction
represent a unique subset of zinc hydrolases that hydrolyze
the b-lactam ring in several antibiotics (Scheme 1). The drug
resistance that results from this hydrolysis is becoming an
Zinc, an essential and the second-most-abundant transition
metal in biology, exerts its biological effect through several
enzymes.[1] Apart from enzymes with one zinc-binding site,
such as carbonic anhydrase and carboxypeptidase A, en-
zymes containing two or three zinc ions at the active sites
are of current particular interest.[2] The most important and
distinguishing features of these zinc enzymes are that the
two metal ions are connected by bridging ligands with
Zn···Zn distances ranging from 3.0 to 3.5 Š.[3] Metallohydro-
lases with dinuclear-zinc active sites perform many impor-
tant biological hydrolytic reactions on a variety of sub-
strates.[1,2] In this regard, metallo-b-lactamases (mbl, class B)
Scheme 1. Bacterial drug resistance: Hydrolysis of b-lactam antibiotics by
metallo-b-lactamases (mbl).
increased problem for the clinical community.[4] These met-
alloenzymes can hydrolyze a wide range of b-lactam sub-
strates, such as cephamycins and imipenem, that are gener-
ally resistant to the serine-containing b-lactamases.[5,6]
Therefore, the clinical application of the entire range of an-
tibiotics is severely compromised in bacteria that produce
metallo-b-lactamases.
[a] A. Tamilselvi, Dr. M. Nethaji, Dr. G. Mugesh
Department of Inorganic and Physical Chemistry
Indian Institute of Science, Bangalore 560 012(India)
Fax : (+91)80-2360-0683/2360-1552
E-mail: mugesh@ipc.iisc.ernet.in
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author: Syntheses of li-
1
Metallo-b-lactamases use one or two zinc ions in their
active sites and are classified into three groups according to
gands, results of H NMR spectroscopic investigations of the hydroly-
sis of penicillin G, and HPLC data for the hydrolysis of oxacillin.
Chem. Eur. J. 2006, 12, 7797 – 7806
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7797
Pagination corrected Oct. 17, 2006

their amino acid sequences.[4e,5a] The largest and most well-
studied group of enzymes that belongs to the subclass B1 in-
cludes BcII from Bacillus cereus, CcrA from Bacteroides fra-
gilis (Figure 1), and IMP-1 from Pseudomonas aeruginosa.
The subclass B2prototype enzyme is CphA from Aeromo-
plex acts as the nucleophile in hydrolysis, and that the car-
boxylate group of these antibiotics is used for favorable po-
sitioning of the substrate and the nucleophile. Meyer and
co-workers, on the other hand, investigated the hydrolytic
cleavage of penicillin G mediated by pyrazolate-based dinu-
clear zinc complexes and showed that the cooperative ef-
fects of the adjacent metal ions might be operative, although
they either enhance or diminish the b-lactamase activity
with respect to free mononuclear zinc.[8d] Another exciting
development in this area is the conversion of glyoxalase II
into metallo-b-lactamase by extensive sequence changes, in-
cluding the deletion and insertion of several structural loops
in the active site. The resulting enzyme, evMBL8 (evolved
metallo-b-lactamase 8), completely loses its original activity
and, instead, catalyzes hydrolysis of a b-lactam substrate, ce-
fotaxime.[9] In this paper, we describe a series of novel
mono- and dinuclear zinc complexes as functional mimics of
mbl entities. In addition, we describe the presence of a free
tertiary amino substituent in close proximity to the zinc-
bound water, which acts as a general base, promoting hy-
drolysis by activating the water molecule.
Results and Discussion
Figure 1. The active site of a) binuclear zinc enzyme from B. fragilis
(PDB code: 1ZNB) and b) mononuclear zinc enzyme from B. cereus
(PDB code: 1BMC).
The ligands HL1–HL3 (Figure 2) used in the present study
were synthesized from p-cresol and p-bromophenol by using
appropriate amines. The amino phenol ligands HL1–HL3
nas hydrophila, which differs
considerably from other en-
zymes in its zinc dependence
and substrate specificity. The
third subclass B3 includes the
L1 enzyme from Stenotropho-
monas maltophilia, which is the
only tetrameric mbl described
so far. The crystal structures of
several mbl entities reveal a
unique ab/ba sandwich fold in
which the catalytic zinc-binding
site lies at the interface be-
tween the two domains. Howev-
er, the role of the additional
zinc in the binuclear enzymes is
still
a
matter of debate, as
mono
ACHTREUNG
display hydrolytic activity com-
parable with that of the binu-
clear enzymes.[7]
Figure 2. Chemical structures of ligands HL1–HL3 and complexes 1–4.
Recently,
a great deal of
effort has been devoted to the development of zinc com-
plexes as mimics of mbl entities.[8] In their pioneering work,
Lippard and co-workers showed that zinc complexes having
mono- and dinuclear active centers can catalyze the hydroly-
sis of penicillin G and nitrocefin.[8b] They also demonstrated
that the bridging hydroxide in the dinuclear zinc(II) com-
were chosen because these dinucleating systems can hold
two zinc(II) ions together, mimicking the zinc environment
at the active sites of metallo-b-lactamases, in which the two
zinc ions are coordinated predominantly by histidine ligands.
Suitably designed phenolate ligands with ethylene diamine
arms strongly favor the formation of dimetallic com-
7798
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7797 – 7806

Metallo-b-Lactamase Mimics
FULL PAPER
plexes.[10–13] Sorrel and co-workers reported some phenolate-
bridged copper dimers as models for the active site of oxi-
dized-hemocyanin derivatives and Karlin and co-workers
utilized such ligands for modeling the copper monooxyge-
nase systems.[10,13] Although there are several other exam-
ples of phenolate-bridged ligands,[14] only few of them have
been used for hydrolytic reactions.[15] Uhlenbrock and Krebs
reported the first example of a dizinc complex having a phe-
noxy bridge as a model for the dinuclear unit in the active
site of phospholipase C.[16] Lippard and co-workers, on the
other hand, reported an interesting class of dinuclear
zinc(II) complexes having two ethylenediamine chelating li-
gands as functional mimics of metallo-b-lactamases.[8b]
dination, but in this case the sixth coordination site is com-
pleted by a water molecule. The Zn···Zn distance of
3.355(7) Š observed in complex 1 is significantly shorter
than that of the binuclear zinc enzyme from B. cereus (3.848
and 4.365 Š),[3a] however, it is comparable with the corre-
sponding distances in other binuclear enzymes, such as S.
maltophilia (3.459 Š) and B. fragilis (3.457 Š).[3b,17] Complex
1 appears to be a good model for the active site of metallo-
b-lactamases, because all three classes of metallo-b-lacta-
mases mentioned above have a water molecule that coordi-
nates to one of the zinc ions.
The synthesis of complex 1 was highly reproducible and
the crystallization in dichloromethane by using different
concentrations of the reaction product invariably afforded
complex 1. Because the stability constant of this complex is
not expected to be very high, the acetates may dissociate
and/or the dinuclear species may become mononuclear in
solution. Therefore, the molecular structure of the complex
in solvents, such as DMSO, in which the hydrolytic reactions
were carried out (see below) may not be identical to that
obtained from dichloromethane. To confirm the stability of
complex 1 in solution, we carried out HPLC experiments in
dichloromethane and in DMSO.
Synthesis of the novel dinuclear zinc complex 1 (Figure 2)
having a coordinated water molecule was achieved in good
yield by treating the sodium salt of 2,6-bis[(dimethylamino)-
methyl]-4-methyl phenol (HL1) with Zn
(OAc)2·6H2O. The
A
crystals suitable for single-crystal X-ray diffraction studies
were obtained from a dichloromethane solution. The X-ray
crystal structure of complex 1 reveals that the coordination
geometry around the Zn1 ion differs from that of the other
zinc ion (Zn2) (Figure 3). The first zinc ion (Zn1) is five-co-
These experiments show that
the chromatograms obtained
from the DMSO solution are
identical to the ones obtained
from the dichloromethane solu-
tion, indicating that complex 1
retains its dinuclear structure in
solution. The peak areas ob-
tained at different time inter-
vals also suggest that the com-
plex stays intact in solution
(Figure S6b, Supporting Infor-
mation).
The other dinuclear zinc
complex
[Zn2(L3)2Cl2]
(2)
(Figure 2) was synthesized by
treating the amino phenol HL3
with zinc chloride in the pres-
ence of triethylamine. The crys-
tals suitable for X-ray analysis
were obtained from a dichloro-
methane solution. In contrast to
complex 1, this compound crys-
tallized as a symmetric dinu-
clear zinc complex with
Zn···Zn distance of 3.223 Š,
which is significantly shorter
a
Figure 3. X-ray crystal structures of 1–4. In 1, the water molecule is hydrogen bonded to one of the acetate li-
gands, whereas in 3 and 4, the water molecules are hydrogen bonded with the tertiary amino substituents. The
hydrogen atoms (except those of water molecules) are omitted for clarity.
than that of complex
1
(Figure 3). The two metal ions
ordinate with four oxygen atoms and a nitrogen atom. The
phenolate oxygen and an acetate ion act as bridging ligands,
whereas the amino nitrogen and an asymmetric acetate
group act as terminal ligands. The second zinc ion (Zn2), on
the other hand, is six-coordinate with the same type of coor-
are held together by two phenolate bridging ligands and
each metal ion is surrounded by two oxygen atoms, two ni-
trogen atoms, and a chloride ion. Notably, the presence of
two ethylenediamine arms instead of just one in the ortho
positions led to the isolation of dinuclear complexes having
Chem. Eur. J. 2006, 12, 7797 – 7806
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7799

G. Mugesh et al.
one phenolate and a hydroxyl, methoxy, or nitrate bridging
ligand.[8b] The Zn···Zn distance observed in complex 2 is
comparable with the corresponding distances in these zinc
complexes, which range from 3.115 to 3.229 Š.[8b] In the syn-
theses of zinc complexes from zinc chloride, the nature of
the ligands appears to have a large effect on the complex
formation. If the reaction of zinc chloride was carried out
with HL1 instead of HL3 (see Figure 3), we isolated a yellow
compound that was a mononuclear species (3). Similarly, the
reaction of HL2 with anhydrous zinc chloride afforded an-
other mononuclear complex (4) in moderate yield. The X-
ray crystal structures of complexes 3 and 4 show trigonal-bi-
pyramidal arrangements of the ligands around the metal
center. Interestingly, in both complexes, one of the coordina-
tion sites around the zinc center is occupied by a water mol-
ecule (Figure 3). All attempts to prepare the corresponding
dinuclear complexes by treating ligands HL1 and HL2 with
zinc chloride were unsuccessful, although these ligands are
known to be dinucleating species.
The dinuclear structures of complexes 1 and 2 and the ob-
servation of coordinated water molecules in complexes 1, 3,
and 4 prompted us to study the hydrolytic behavior of these
complexes. The b-lactamase activity of the metalloenzyme
from Bacillus cereus and complexes 1–4 was studied by em-
1
ploying H NMR spectroscopy and the reverse-phase HPLC
method. Interestingly, complex 1 exhibited high b-lactamase
activity by hydrolyzing the antibiotic substrate penicillin G
(Scheme 1, R = Bz) at room temperature and a complete
hydrolysis took place within 3 h, upon using 50 mm of the
catalyst and 50 mm of the substrate (Figure S2, Supporting
Information). Similarly, this complex was efficient in the hy-
drolysis of oxacillin, a penicillin analogue with high resist-
ance to most of the serine-containing b-lactamases and,
therefore, more stable than penicillin G toward hydrolysis.
The hydrolyzed b-lactam products in these reactions were
identical to the enzymatic reaction products, as evidenced
from our 1H NMR spectroscopic and HPLC analyses
(Figure 4). The NH group of penicillin G and oxacillin was a
Figure 4. The NH and CH regions from the 1H NMR spectra of a) oxa-
cillin, b) an equimolar mixture of oxacillin and 1, c) an equimolar mixture
of oxacillin and 3, and d) an equimolar mixture of oxacillin and 4. The
spectra were recorded in [D6]DMSO at RT.
1
useful probe for our H NMR spectroscopic investigations.
Treatment of complex 1 with penicillin G or oxacillin result-
ed in a complete change in the characteristic b-lactam sig-
nals, due to the hydrolytic cleavage of the four-membered
ring. The nature of the hydrolyzed products was confirmed
by comparing the NH and CH signals with those of the
products obtained from the hydrolysis of penicillin G or ox-
acillin by the natural b-lactamase enzyme.
To find out whether complex 1 can act as a catalyst in the
hydrolysis reaction, we undertook a detailed kinetic analysis.
Interestingly, complex 1 was highly active at concentrations
as low as 0.3 mol% (Figure S18 v–x, Supporting Informa-
tion). The plots of percentage of oxacillin versus time for
hydrolysis in the presence of 1 and some related compounds
are given in Figure 5, and the corresponding t50 values (the
time required for hydrolysis of 50% of the substrate) for ox-
acillin and penicillin G are summarized in Table 1. These
data indicate that compound 1 rapidly hydrolyzes oxacillin
and penicillin G with t50 values of 97 and 77 min, respective-
Figure 5. Hydrolysis of oxacillin (1.06”10À3 m) by a) HL1, b) zinc acetate,
c) 4, d) 3, e) 2, f) 1 (each 2.1”10À5 m) in DMSO/HEPES buffer (1:9),
pH 7.5. The reaction was monitored by reverse-phase HPLC and the
amount of oxacillin hydrolyzed at any given time was calculated from the
calibration plot.
ly, and these values are much lower than that of the control
reaction (t50 >6300 min for oxacillin and >650 min for peni-
cillin G). Furthermore, a complete hydrolysis of the sub-
strates (100% hydrolysis) was observed with complex 1. In
contrast, a complete hydrolysis of the substrate was not ob-
served with Zn
though Zn(OAc)2·2H2O exhibited a significant initial reac-
tion rate (Table 1, entry h) relative to the experiments not
A
AHCTREUNG
7800
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7797 – 7806

Metallo-b-Lactamase Mimics
FULL PAPER
Table 1. The t50 values and initial rates (v0) of hydrolyses of oxacillin and
been made for the hydroxylating activity of the diiron meth-
ane monooxygenase (MMO) enzyme in which the terminal
water molecule attached to one of the iron ions is activated
by Glu residues.[20] In agreement with this, the binuclear zinc
complex 2 with no coordinated water or acetate ligands
(Figure 3) shows much lower b-lactamase activity than that
of 1. The t50 values observed for complex 2 in the hydrolysis
of oxacillin and penicillin G are three and two times, respec-
tively, higher than that for complex 1. A similar trend was
observed in the initial rate of hydrolysis (Table 1).
penicillin G by BcII, HL1, Zn
(OAc)2·2H2O, and complexes 1–4 at RT.
A
Entry Catalyst
Oxacillin
Penicillin G
t50
v0
t50
v0
[e]
[e]
A
A
]
A
A
]
a
b
c
d
e
f
control
>6300[b]
229
0.76
>650[f]
–
0.05
–
0.25
15.53
6.86
6.57
3.14
1.02
BcII
12.67
0.32
13.83
5.09
6.11
3.29
0.81
HL1
>1940[c]
97
>650[g]
77
1
2
3
4
289
353
136
214
g
h
549
563
To probe the role of the second metal ion in complex 1,
we carried out detailed hydrolysis experiments with the
mononuclear zinc complexes 3 and 4 that have a water mol-
ecule as one of the ligands. The hydrolysis of penicillin G
and oxacillin was monitored by 1H NMR and/or HPLC
methods. Interestingly, these complexes also hydrolyzed
penicillin G (1H NMR) and oxacillin (1H NMR and HPLC)
(Figures 4 and 5; Table 1), although the activity was less
than that of the binuclear zinc complex 1. The t50 values for
the hydrolysis of oxacillin by 3 (353 min) and 4 (549 min)
are higher than that of 1, however, these values are very
much lower than that of the control reaction (>6300 min).
The initial rates (v0) for the hydrolysis of oxacillin and peni-
cillin G by 3 and 4 are higher than that of the control rate
and also significantly higher than that of zinc acetate
(Table 1). These observations support the previous results
by Lippard and co-workers that the second metal ion is not
crucial for b-lactamase activity.[8b] The greater activity of
complexes 3 and 4 relative to that of free zinc ions is proba-
bly due to the presence of the coordinated water molecules
that are involved in strong hydrogen bonding with the unco-
ordinated tertiary amino groups (Figure 3). The N···HO hy-
drogen-bonding distances in complexes 3 and 4 were 1.909
and 2.000 Š, respectively. These amino groups would cer-
tainly activate the water molecules to enhance the hydroly-
sis reaction.
The catalytic mechanism of Zn-containing enzymes re-
quiring one metal ion for activity may become somewhat
more efficient by acquisition of cocatalytic sites with two
zinc ions in close proximity acting as a unit center.[21] How-
ever, the studies on the B. fragilis enzyme showed that the
catalytic efficiency of the dinuclear enzyme is only marginal-
ly superior to that of its mononuclear counterpart with some
substrates, and is even lower with other ones.[3d] Recent
studies on functional analysis of the active-site residues
reveal that two amino acids residues, Asp90 and His210,
form a catalytic center with the zinc-bound water molecule,
in which Asp90 acts as a general base in the enzymatic proc-
ess. Further studies reveal that the proton-transfer process[22]
from zinc-bound water to His210 readily occurs through a
water-assisted pathway with an estimated barrier of around
8 kcalmolÀ1.[19a,b] The proton abstraction by the histidine res-
idue converts the water molecule into a more reactive hy-
droxo species, facilitating hydrolysis of the b-lactam ring.
Based on these observations, it is clear that the uncoordinat-
ed amino substituent in complexes 3 and 4 acts as a general
base in the reaction. This further supports the assumption
Zn
A
>750[d]
>700[h]
[a] Time required for 50% hydrolysis of substrate (t50) was determined
by monitoring the decrease in peak area due to oxacillin/penicillin G. [ox-
acillin]: 1.06”10À3 m; [penicillin G]: 1.00”10À3 m; [catalyst]: 2.12”10À5 m;
in DMSO/HEPES buffer (1:9), pH 7.5; [BcII, class B enzyme from B.
cereus]: 44.2n m in HEPES buffer, pH 7.5. [b] After 6300 min, only 36%
hydrolysis was observed. [c] After 1940 min, only 24% hydrolysis was ob-
served. [d] After 750 min, only 36% hydrolysis was observed. [e] Calcu-
lated from the initial 5–10% of reaction by monitoring the decrease in
peak area due to oxacillin. [f] After 650 min, only 4% hydrolysis was ob-
served. [g] After 650 min, only 10% hydrolysis was observed. [h] After
700 min, only 26% hydrolysis was observed.
involving metal ions. Similarly, the free ligand (HL1) did not
show any noticeable activity, as the initial rate observed in
the presence of HL1 was almost identical to (penicillin G,
Figure S17, Supporting Information) or even lower than (ox-
acillin) that of the control rate. With oxacillin as the sub-
strate, only 24% hydrolysis was observed after 1940 min in
the presence of HL1 (Table 1, entry c) and 36% conversion
was observed after 750 min in the presence of Zn-
(OAc)2·2H2O (Table 1, entry h). The natural enzyme (BcII,
44.2n m), on the other hand, completely and effectively hy-
drolyzed oxacillin with a t50 value of 229 min. These observa-
tions also suggest that the simple metal salts are dissociated
in solution to give free zinc ions that may be inhibited by
the products, whereas the preorganized dinuclear zinc
center in the natural enzyme and in complex 1 provides gen-
eral base and single-site reactivity.
The remarkable catalytic activity of complex 1 can be as-
cribed to the tightly bound dinuclear zinc center and also to
presence of an activated zinc-bound water molecule. It has
been observed that the stability of the dizinc complexes in
water is crucial for the hydrolytic activity of synthetic com-
pounds. For example, certain active dinuclear zinc com-
plexes that exhibit aminopeptidase activity are deactivated
in the presence of water.[18] Therefore, the stability of the di-
nuclear metal center in complex 1 may account for its high
reactivity. Furthermore, the analysis of the coordination en-
vironment around the zinc ions in complex 1 reveals that
the water molecule coordinated to the Zn2center is hydro-
gen bonded to the oxygen atom of the acetate ligand attach-
À
ed to Zn1 (O H···O: 1.944 Š) (Figure 3), which may acti-
vate the water molecule toward hydrolysis. This observation
is supported by the fact that one of the amino acid residues
(Asp90) acts as the general base to abstract the proton from
water and assist in hydrolysis.[19] Similar observations have
Chem. Eur. J. 2006, 12, 7797 – 7806
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7801

G. Mugesh et al.
that residues Asp90 and His210, which are involved in coor-
dination with the second zinc in B. cereus (Figure 6), would
play the role of the second zinc in mononuclear b-lactamas-
es by activating the water molecules (Figure 1b).
Therefore, the initial rate is first order with respect to the
complex concentration and hydrolysis may take place in a
bimolecular fashion, analogous to the hydrolysis of penicillin
catalyzed by [Zn
(cyclen)(OH)]+ in aqueous solution.[8a]
A
However, our further studies indicate that the substrate
binds more tightly to the complex, as evidenced from the
low Michaelis constant (Km) (see below), indicating an ini-
tial binding of the substrate, followed by intramolecular hy-
drolysis by the coordinated water/hydroxide.
As the catalyst (1) concentration was kept constant and
the substrate (oxacillin) concentration was increased, a
rapid increase in rate was observed in the initial stages, how-
ever, as the substrate concentration was increased further,
the rate became constant, indicating that hydrolysis of oxa-
cillin by the zinc complexes follows Michaelis–Menten ki-
netics. We next set out to determine kinetic parameters,
such as the maximum velocity (Vmax), Michaelis constant
(Km), turnover number (kcat), and catalytic efficiency (h), for
the hydrolysis reaction by carrying out several HPLC ex-
periments at various concentrations of oxacillin. Double-re-
ciprocal plots (Lineweaver–Burk plots) of initial rate versus
substrate concentration yielded families of linear lines for
all the complexes (Figure 8). The lines correspond to differ-
Figure 6. Active site of binuclear zinc enzyme from B. cereus (BcII, PDB
code: 1BC2).
To understand the catalytic hydrolysis of b-lactam sub-
strates by the synthetic compounds, and the effect of catalyst
and substrate concentration on the reaction rates, we carried
out further experiments by using oxacillin as the substrate.
The rate of increase in the product formation was monitored
by the reverse-phase HPLC method and the chromatograms
were obtained at 254 nm, at which the hydrolyzed product
showed a maximum absorbance. These measurements were
recorded at various concentrations of complex ranging from
3.3 to 20 mm (Table 1) by keeping the substrate concentra-
tion constant. The initial rates were obtained graphically
from the time course of each experiment. These results
show that complex 1 is active in the hydrolysis of oxacillin
at concentrations as low as 0.3 mol%. In other words, com-
plex
1 catalyzes the complete hydrolysis of at least
300 equivalents of oxacillin substrate. Upon performing
these experiments at different concentrations of complex 1,
a steady increase in the rate was observed as concentration
increased. The initial rates (v0) obtained at various concen-
trations of 1 are summarized in Table 1. The plot of initial
rates against the catalytic concentration gave a straight line,
indicating that b-lactamase activity is linearly proportional
to the concentration of catalyst (Figure 7).
Figure 8. Lineweaver–Burk plots. Effect of substrate (oxacillin) concen-
tration [S] on the lactamase activity of a) BcII, b) 1, c) 3, d) 2, and e) 4.
ent catalysts and indicate that the rates increase linearly as
concentration of the catalyst increases. The kinetic parame-
ters obtained at several oxacillin concentrations are summar-
ized in Table 2.
Interestingly, the Michaelis constants (Km) for the dinu-
clear complexes 1 (8.94 mm) and 2 (3.43 mm) were much
lower than those for the mononuclear complexes
3
(34.30 mm) and 4 (11.56 mm) under similar conditions, indi-
cating that the binding of the substrate to the dinuclear
complexes is much stronger than that to the mononuclear
complexes. Furthermore, the Km value for complex 1 in the
presence of oxacillin is very similar to that of the dinuclear
zinc enzymes BcII (6.97 mm) and IMP-1 (7.90 mm), but much
lower than that of evMBL8 (Table 2), indicating that the
binding of substrate to the dinuclear center of 1 is almost
the same as that to the natural enzymes, but much stronger
Figure 7. Effect of catalyst concentration on the hydrolysis of oxacillin by
1 (values taken from Table S20a). Initial rates were determined by moni-
toring the increase in peak area due to formation of hydrolyzed oxacillin
turnover product, and were corrected for their background rates; [oxacil-
lin]=1.06”10À3 m in DMSO/HEPES buffer (1:9), pH 7.5.
7802
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7797 – 7806

Metallo-b-Lactamase Mimics
FULL PAPER
Table 2. The maximum rate (Vmax), Michaelis constant (Km), turnover
number (kcat), and catalytic efficiency (h) for the hydrolysis of oxacillin
(1–4, BcII), and cefotaxime (evMBL8, IMP-1).
pected, the catalytic activity was completely abolished at
lower pH values and a rapid increase in the rate of hydroly-
sis was observed for complexes 1 and 3 at pH values above
7.0 (Figure 9). This is consistent with the observations of
Lippard and co-workers that the rate-limiting nucleophilic
Complex
Vmax
[mmminÀ1
Km
[mm]
kcat
h
G
]
Vmax
/
N
kcat/Km
[minÀ1
[mminÀ1
]
1
2
3
4
28.36
9.08
83.3234.30
16.10
–
–
8.94
3.43
8.33
11.56
229.00
7.90
2.83
0.91
.43”10
1.61
2.52
378.00
717.20
3.17”105
2.64”105
5
2
1.39”105
0.11”105
4.78”107
evMBL8[a]
IMP-1[a]
BcII[b]
31.64
6.97
10.29”107
[a] ¾pparent kinetic parameters are taken from ref. [9]. [b] Initial con-
centration of BcII was fixed at 44.2n m. All parameters listed are based
on the initial rates determined at 258C.
than that to the engineered enzyme evMBL8. This is in ac-
cordance with the report of Rasia and Vila that states that
the Km values for the dinuclear native enzyme and their mu-
tants from B. cereus are much lower than that of their
Figure 9. Effect of pH on the rate of hydrolysis of oxacillin by 1 and 3.
mono
nuclear analogues.[23] Because the substrate binds to
A
the dinuclear complexes more tightly than to the mononu-
clear ones (see below), saturation kinetics were easily ob-
served for the dinuclear complexes. These observations sug-
gest that the second zinc in the enzyme and also in the com-
plexes plays some role in stabilizing the catalyst–substrate
intermediate. These data also suggest that the dinuclear cen-
ters in complexes 1 and 2 stay intact in the hydrolysis, which
is important for the binding of substrate, as suggested by
Lippard and co-workers for the hydrolysis of nitrocefin.[8c]
In contrast, the kcat value of the mononuclear zinc com-
plex 3 is increased by 66% relative to that of the dinuclear
complex 1. This is due to the Vmax value observed for com-
plex 3 being higher than that of 1. This indicates that the un-
coordinated amino substituents in 3 may play important
roles in hydrolysis. However, the overall catalytic efficien-
cies (h) derived from the kcat and Km values confirm that
complex 1 is superior to other di- and mononuclear zinc
complexes (2–4) in the present study. Crucially, the catalytic
efficiency (h) of this particular complex for oxacillin (3.17”
105 mÀ1 minÀ1) is almost 30-times higher than that of
evMBL8 for cefotaxime (0.11”105 mÀ1 minÀ1),[9] and only
150- and 325-times lower than that of the dinuclear zinc en-
zymes IMP-1 and BcII, respectively (Table 2). These obser-
vations suggest that complex 1 has acquired a well-defined
dinuclear metal coordination and a substrate-binding pocket
for the metallo-b-lactamase activity. These data also suggest
that the metal complexes exhibiting lower Km and higher
Vmax are better catalysts for hydrolysis than the ones that ex-
hibit higher Km and lower Vmax or higher Km and Vmax
values. According to this, complex 1 is considered to be the
most effective catalyst and complex 4 is considered to be
the least effective catalyst in the present study, which is con-
sistent with the t50 values for the hydrolysis of oxacillin and
penicillin G by complexes 1 and 4 (Table 1).
attack of the hydroxo species on the substrate, followed by
fast protonation of the intermediate, leads to the formation
of the hydrolysis product.[24] Although the rate of reaction
became constant at pH 8.5 for complex 1, no such behavior
was observed for complex 3. In the case of 3, a continuous
increase in the rate was observed until pH 11.0, confirming
the participation of the free tertiary amino substituents in
hydrolysis. This is in agreement with the report of Bounaga
and co-workers that protonation of the catalytically impor-
tant Asp120 residue contributes to enzyme inactivation at
acidic pH.[25] These data also suggest that the coordinated
substrate may be attacked by an external hydroxide, leading
to hydrolysis of the substrate. Rasia and Vila showed that
the reversible inactivation of the enzyme at low pH may
also be due to dissociation of the two zinc(II) equivalents.[23]
The pH profile indicated that a second zinc(II) equivalent is
not needed for nucleophile activation. Instead, the second
zinc(II) may act to properly anchor Asp120, which ultimate-
ly orients the attacking nucleophile in dinuclear enzymes.
They also suggested that the Arg121 residue may fulfil the
role of the second zinc in the mononuclear enzymes.
Conclusion
We have shown that suitably substituted mono- and dinu-
clear zinc complexes can act as catalysts for the hydrolysis
of b-lactam substrates, mimicking the properties of metallo-
b-lactamases. This study suggests that the high reactivity of
the complexes is due to the presence of zinc-bound water
molecules that are activated by acetate or amino substitu-
ents. Our results support the assumption that the second
zinc in the dinuclear enzymes does not participate directly
in catalysis, but may orient the substrates for hydrolysis, and
the basic amino acid residues, such as Asp and His, may ac-
To understand the effect of pH on the reaction rates, the
initial rates were determined at various pH values. As ex-
Chem. Eur. J. 2006, 12, 7797 – 7806
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7803

G. Mugesh et al.
(m), 2732 (w), 2490 (w), 1612 (vs), 1494 (s), 1463 (m), 1417 (w), 1391 (w),
1345 (w), 1266 (vs), 1228 (m), 1168 (m), 1126 (m), 1104 (m), 1061 (m),
1010 (m), 975 (w), 939 (s), 835 (s), 792(s), 732(m), 757 (m), 699 (w),
672cm À1 (w); MS (ESI): m/z: 643.2[Zn 2(L3)2Cl2]+; elemental analysis
calcd (%) for C26H42N4O2Cl2Zn2 (644.3): C 48.47, H 6.57, N 8.70; found:
C 44.86, H 7.03, N 8.28.
tivate the zinc-bound water molecules, fulfilling the role of
the second zinc in the mononuclear enzymes. However, our
HPLC results from the hydrolysis of oxacillin and penicil-
lin G suggest that the second metal ion is required for a
quantitative conversion of the substrates to the hydrolyzed
products.
Synthesis of [Zn(L1)2(H2O)] (3): The ligand HL1 (0.287 g, 1 mmol) in di-
A
chloromethane (6 mL) together with 60% NaH (0.036 g, 0.9 mmol) in
hexane (25 mL) were mixed and stirred at RT for 1 h to obtain the corre-
sponding sodium phenolate (NaL1). Anhydrous ZnCl2 (0.137 g,
1.5 mmol) was added and the stirring was continued for an additional
1 h. The yellow turbid solution was filtered off to afford a clear solution,
from which the yellow crystals suitable for X-ray analysis were obtained.
Yield: 0.185 g (35%). 1H NMR (CDCl3): d=2.33 (s, 6H), 2.29 (s, 24H),
3.50 (s, 8H), 6.83 ppm (s, 4H); IR (neat): n˜ =3400 (br), 2939 (m), 2853
Experimental Section
General: Penicillin G (sodium salt) and oxacillin (sodium salt) were ob-
tained from Fluka; captopril and penicillanase (b-lactamase II, BCII)
from Bacillus cereus were obtained from Sigma. HEPES sodium salt,
sodium hydride, deuterated solvents D2O, CD3OD, [D6]DMSO, and
CDCl3 were obtained from Aldrich. N,N-dimethylamine, N,N,N’-tri-
(m), 2817 (m), 2774 (m), 1612 (w), 1473 (m), 1355 (w), 1321 (m), 1258
À1
(w), 1033 (br, s), 871 (w), 841 (w), 792(w), 606 (w), 459 cm
(s); MS
(ESI): m/z: 508 [Zn(L1)2]+; elemental analysis calcd (%) for
C26H44N4O3Zn (526.08): C 59.36, H 8.43, N 10.65; found: C 58.89, H 8.23,
N 10.25.
G
A
cresol, p-bromophenol, and 37% formaldehyde solution were obtained
from other commercial sources and were used as received. HPLC-grade
solvents for the kinetic experiments were obtained from Merck.
Synthesis of [Zn(L2)2(H2O)] (4): The ligand HL2 (0.287 g, 1 mmol) in di-
G
All chemical reactions were carried out under nitrogen or argon by using
standard vacuum-line techniques. Solvents were purified by standard pro-
cedures and were freshly distilled prior to use. 1H (400 MHz) and 13C
(100 MHz) NMR spectra were obtained by using a Bruker Avance 400
NMR spectrometer. Chemical shifts are cited in ppm with respect to
SiMe4 as internal (1H and 13C) standard. Infrared spectra were obtained
as neat films by using a JASCO FTIR-410 spectrometer in the 4000–
400 cmÀ1 regions. A Perkin–Elmer Lambda 5 UV/Vis spectrophotometer
was used to measure the electronic absorption spectra. Mass spectral
studies were carried out by using either a Q-TOF micro mass spectrome-
ter with ESI-MS mode analysis or a FAB mass spectrometer. In the case
of isotopic patterns, the value given is for the most intense peak. The syn-
theses of ligands HL1, HL2, and HL3 were carried out by using standard
protocols and are given in the Supporting Information.
chloromethane (5 mL) together with 60% NaH (0.036 g, 0.9 mmol) in
hexane (20 mL) were mixed and stirred at RT for 1 h to obtain the corre-
sponding sodium phenolate (NaL2). The resulting solution was treated
with anhydrous ZnCl2 (0.137 g, 1.5 mmol) and the stirring was continued
for an additional 1 h. The resulting turbid solution was filtered off to
afford a clear solution. The slow evaporation of the solvent afforded
compound 4 as pale yellow crystals that were collected and dried. Yield:
0.269 g (41%). 1H NMR (CDCl3): d=2.29 (br, 24H), 3.50 (s, 8H),
7.14 ppm (s, 4H); 13C NMR (CDCl3): d=165.3, 155.1, 130.4, 124.8, 109.4,
101.8, 59.2, 44.2 ppm; IR (neat): n˜ =2976 (m), 2946 (m), 2853 (m), 2776
(m), 2341 (w), 1580 (m), 1456 (s), 1421 (s), 1363 (m), 1326 (m), 1289 (w),
1176 (m), 1150 (m), 998 (s), 982(m), 843 (m), 742 (m), 624 (m), 500 (w),
451 cmÀ1 (w); MS (FAB): m/z: 637 [Zn(L2)2]+; elemental analysis calcd
(%) for C24H38Br2N4O3Zn (655.78): C 43.96, H 5.84, N 8.54; found: C
44.22, H 5.99, N 8.39.
Synthesis of [Zn2L1
A
G
N
1 mmol) in dichloromethane (5 mL) together with 60% NaH (0.036 g,
0.9 mmol) in hexane (25 mL) were mixed and stirred at RT for 1 h to
X-ray crystallography: X-ray crystallographic studies were carried out by
obtain the corresponding sodium phenolate (NaL1). Zn
(OAc)2·2H2O
U
using
a Bruker CCD diffractometer with graphite-monochromatized
MoKa radiation (l=0.71073 Š) controlled by a Pentium-based PC run-
ning the SMART software package.[26] Single crystals were mounted at
RT on the ends of glass fibers and data were collected at 298 K. Intensity
data were measured in frames with increasing w (width of 0.38 per
frame) at a scan speed of 18 s per frame and the SMART and SAINT
software were used for data acquisition and data extraction, respectively.
The structures were solved and refined by using the SHELXTL software
package.[27] In general, all non-hydrogen atoms were refined anisotropi-
cally. Hydrogen atoms were assigned idealized locations. Empirical ab-
sorption corrections were applied to all structures by using SADABS.[28]
The perspective views of the complexes were obtained by using
PLATON[29] or ORTEP.[30] The coordination environments around the
metal centers are given in Figure 3. CCDC-294711 (1), CCDC-294712
(2), CCDC-294713 (3), CCDC-294714 (4) contain the supplementary
crystallographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data request/cif.
1H NMR study on the hydrolysis of penicillin G and oxacillin by zinc
complexes: The hydrolysis of penicillin G (Figures S1–4, Supporting In-
formation) and oxacillin (Figure 4) was monitored by conducting
1H NMR spectroscopy. All NMR experiments were carried out at 293 K.
The chemical shifts are given in ppm relative to [D6]DMSO as an inter-
nal reference. In each experiment, the test solution contained 0.05 mm of
complex and 0.05 mm of either penicillin G sodium salt or oxacillin
sodium salt. Pure [D6]DMSO was used as the solvent in experiments in-
vestigating the binding of penicillin G with the complex.
(0.329 g, 1.5 mmol) was then added and the stirring was continued for an
additional 1 h. This resulted in a turbid solution from which the zinc com-
plex [Zn2L1
A
G
N
product was dried under vacuum to afford a white crystalline solid. The
solid product was dissolved in dichloromethane and was kept for crystal-
lization. Upon standing at RT, the product crystallized out as white nee-
dles suitable for X-ray analysis. Yield: 0.467 g (85%). 1H NMR (CDCl3):
d=6.80 (s, 2H), 3.73 (brs, 4H), 2.45 (brs, 12H), 2.38 (brs, 2H), 2.27 (s,
3H), 2.02 ppm (s, 9H); 1H NMR ([D6]DMSO): d=6.81 (s, 2H), 3.51
(brs, 4H), 2.20 (brs, 12H), 2.15 (s, 3H), 1.83 ppm (s, 9H); 13C NMR
(CDCl3): d=20.3, 22.7, 46.2, 63.2, 122.4, 123.3, 124.7, 131.6, 132.5, 158.6,
180.3 ppm; IR (neat): n˜ =3444 (s), 3054 (w), 2975 (m), 2918 (m), 2849
(m), 2796 (m), 1744 (w), 1594 (vs), 1480 (m), 1456 (w), 1394 (m), 1338
(m), 1308 (m), 1282(w), 1151 (m), 1105 (w), 1047 (m), 1027 (s), 972
(m),932(w), 875 (m), 799 (m), 767 (m), 734 (w), 678 (s), 619 (w), 566 (s),
405 cmÀ1 (m); MS (FAB): m/z: 469 [Zn2L1(OAc)2]+; elemental analysis
N
calcd (%) for C19H32N2O8Zn2 (547.25): C 41.70, H 5.89, N 5.12; found: C
41.94, H 6.11, N 5.22.
Synthesis of [Zn2(L3)2Cl2] (2): Anhydrous ZnCl2 (0.136 g, 1.0 mmol) in
ethanol (10 mL) was added to a solution of the ligand HL3 (0.222 g,
1.0 mmol) in ethanol (10 mL), and the reaction mixture was stirred for
30 min at RT. Triethylamine (257 ml, 2mmol) was added to the reaction
mixture and the stirring was continued for an additional 1 h. The result-
ing solution was concentrated to dryness under vacuum to afford a white
solid. This was dissolved in dichloromethane and the mixture was al-
lowed to evaporate slowly to obtain the compound 2 as white crystals.
Yield: 0.346 g (54%). 1H NMR (CDCl3): d=2.24 (s, 3H), 2.30 (s, 3H),
2.37 (s, 6H), 2.66 (s, 4H), 3.85 (s, 2H), 6.74 (d, J=8.4 Hz, 1H), 6.78 (s,
1H), 6.95 ppm (d, 2H); IR (neat): n˜ =3798 (w), 3002 (m), 2981 (m), 2921
Free penicillin G: 1H NMR ([D6]DMSO): d=8.84 (d, 1H; C(O)N-
amide), 7.33 (m, 5H; aromatic protons), 5.32ppm (brs, 2H).
7804
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7797 – 7806

Metallo-b-Lactamase Mimics
FULL PAPER
Penicillin G in the presence of complex 1: 1H NMR ([D6]DMSO): d=
8.51 (1H; C(O)N-amide), 7.75 (m, 5H; aromatic protons) 4.47 (1H),
4.72ppm (1H).
Penicillin G in the presence of complex 3: 1H NMR ([D6]DMSO): d=
8.31 (brd, 1H; C(O)N-amide), 7.22 (m, 5H; aromatic protons), 4.67 ppm
(brd, 2H).
Acknowledgements
This study was supported by the Department of Science and Technology
(DST) and the Council for Scientific and Industrial Research (CSIR),
New Delhi, India. A.T. acknowledges the University Grants Commission
(UGC), New Delhi, for a fellowship.
Penicillin G in the presence of complex 4: 1H NMR ([D6]DMSO): d=
8.45 (brd, 1H; C(O)N-amide), 7.18 (brs, 5H; aromatic protons),
4.51 ppm (brd, 2H).
[1] a) D. E. Wilcox, Chem. Rev. 1996, 96, 2435–2458; b) W. N. Lips-
comb, N. Sträter, Chem. Rev. 1996, 96, 2375–2433.
Free oxacillin: 1H NMR ([D6]DMSO): d=9.07 (d, 1H; C(O)N-amide),
[2] a) N. Sträter, W. N. Lipscomb, T. Klabunde, B. Krebs, Angew. Chem.
1996, 108, 2158–2191; Angew. Chem. Int. Ed. Engl. 1996, 35, 2 02 5–
2055; b) H. Steinhagen, G. Helmchem, Angew. Chem. 1996, 108,
2489–2492; Angew. Chem. Int. Ed. Engl. 1996, 35, 2339–2342.
[3] a) J. A. Cricco, A. J. Vila, Curr. Pharm. Des. 1999, 5, 915–927;
b) J. A. Cricco, E. G. Orellano, R. M. Rasia, E. A. Ceccarelli, A. J.
Vila, Coord. Chem. Rev. 1999, 190–192, 519–535; c) Z. Wang, W.
Fast, A. M. Valentine, S. J. Benkovic, Curr. Opin. Chem. Biol. 1999,
3, 614–622; d) Z. Wang, S. J. Benkovic, J. Biol. Chem. 1998, 273,
22402–22408.
[4] a) N. Sträter, W. N. Lipscomb, Biochemistry 1995, 34, 14792–14800;
b) B. Chevrier, C. Schalk, H. DꢁOrchymont, J. M. Rondeau, D.
Moras, C. Tarnus, Structure 1994, 2, 283–291; c) J. L. Vanhooke,
M. M. Benning, F. M. Raushel, H. M. Holden, Biochemistry 1996,
35, 6020–6025; d) E. Hough, L. K. Hansen, B. Birknes, K. Jynge, S.
Hansen, A. Hordvik, C. Little, E. Dodson, Z. Derewenda, Nature
1989, 338, 357–360; e) A. Volbeda, A. Lahm, F. Sakiyama, D. Suck,
EMBO J. 1991, 10, 1607–1618.
[5] a) K. Bush, Clin. Infect. Dis. 1998, 27, S48–S53; b) D. M. Livermore,
J. Antimicrob. Chemother. 1998, 41, 25–41.
[6] H. Kurosaki, Y. Yamaguchi, T. Higashi, K. Soga, S. Matsueda, H.
Yumoto, S. Misumi, Y. Yamagata, Y. Arakawa, M. Goto, Angew.
Chem. 2005, 117, 3929–3932; Angew. Chem. Int. Ed. 2005, 44, 3861–
3864.
[7] R. Paul-Soto, M. Herµndez-Valladares, M. Galleni, R. Bauer, M.
Zeppezauer, J.-M. FrØre, H. W. Adolph, FEBS Lett. 1998, 438, 137–
140.
[8] a) T. Koike, M. Takamura, E. Kimura, J. Am. Chem. Soc. 1994, 116,
8443–8449; b) N. V. Kaminskaia, B. Spingler, S. J. Lippard, J. Am.
Chem. Soc. 2000, 122, 6411–6422; c) N. V. Kaminskaia, B. Spingler,
S. J. Lippard, J. Am. Chem. Soc. 2001, 123, 6555–6563; d) B. Bauer-
Siebenlist, S. Dechert, F. Meyer, Chem. Eur. J. 2005, 11, 5343–5352.
[9] H. Park, S. Nam, J. Lee, C. Yoon, B. Mannervik, S. J. Benkovic, H.
Kim, Science 2006, 311, 535–538.
[10] T. N. Sorrell, C. J. OꢁConnor, O. P. Anderson, J. H. Reibenspies, J.
Am. Chem. Soc. 1985, 107, 4199–4206.
[11] T. Mallah, M.-L. Boillot, O. Kahn, J. Gouteron, S. Jeannin, Y. Jean-
nin, Inorg. Chem. 1986, 25, 3058–3065.
[12] E. E. Eduok, C. J. OꢁConnor, Inorg. Chim. Acta 1984, 88, 229–233.
[13] M. S. Nasir, B. I. Cohen, K. D. Karlin, J. Am. Chem. Soc. 1992, 114,
2482–2494.
[14] a) B.-H. Ye, X.-Y. Li, I. D. Williams, X.-M. Chen, Inorg. Chem.
2002, 41, 6426–6431; b) A. Erxleben, Inorg. Chem. 2001, 40, 208–
213; c) C. K. Williams, N. R. Brooks, M. A. Hillmyer, W. B. Tolman,
Chem. Commun. 2002, 2132–2133.
[15] a) J. Chen, X. Wang, Y. Zhu, J. Lin, X. Yang, Y. Li, Y. Lu, Z. Guo,
Inorg. Chem. 2005, 44, 3422–3430; b) A. Horn, Jr., I. Vencato, A. J.
Bortoluzzi, R. Hçrner, R. A. N. Silva, B. Spoganicz, V. Drago, H.
Terenzi, M. C. B. de Oliveira, R. Werner, W. Haase, A. Neves, Inorg.
Chim. Acta 2005, 358, 339–351; c) H. Sakiyama, K. Ono, T. Suzuki,
K. Tone, T. Ueno, Y. Nishida, Inorg. Chim. Acta 2005, 358, 372–374.
[16] S. Uhlenbrock, B. Krebs, Angew. Chem. 1992, 104, 1631–1632;
Angew. Chem. Int. Ed. Engl. 1992, 31, 1647–1648.
[17] N. O. Concha, B. A. Rasmussen, K. Bush, O. Herzberg, Structure
1996, 4, 823–836.
[18] H. Sakiyama, R. Mochizuki, A. Sugawara, M. Sakamoto, Y. Nishida,
M. J. Yamasaki, Chem. Soc. Dalton Trans. 1999, 997–1000.
[19] a) D. Suµrez, Jr., K. M. Merz, J. Am. Chem. Soc. 2001, 123, 3759–
3770; b) N. Dꢂaz, D. Suµrez, Jr., K. M. Merz, J. Am. Chem. Soc.
7.52(m, 5H; aromatic protons), 5.41 ppm (brs, 2H).
Oxacillin in the presence of complex 1: 1H NMR ([D6]DMSO): d=8.65
(1H; C(O)N-amide), 7.63 (brs, 2H; aromatic protons), 7.42 (brs, 3H; ar-
omatic protons), 4.81 (1H), 4.59 ppm (1H).
Oxacillin in the presence of complex 3: 1H NMR ([D6]DMSO): d=8.71
(brd, 1H; C(O)N-amide), 7.64 (brs, 2H; aromatic protons), 7.42 (brs,
3H; aromatic protons), 4.65 ppm (brd, 2H).
Oxacillin in the presence of complex 4: 1H NMR ([D6]DMSO): d=8.70
(brd, 1H; C(O)N-amide), 7.63 (brs, 2H; aromatic protons), 7.42 (brs,
3H; aromatic protons), 4.68 ppm (brd, 2H).
Hydrolysis of penicillin G and oxacillin by b-lactamases: The hydrolyzed
products from the reactions of penicillin G and oxacillin with test com-
pounds were identical to those obtained by using the natural enzymes.
For a comparison, both the binuclear metalloenzyme (BcII) and the
serine-containing enzyme from Bacillus cereus were employed. The reac-
tions were carried out in HEPES buffer at pH 7.5. In a typical experi-
ment, 50 mg of oxacillin was dissolved in 10 mL of HEPES buffer and
this solution was treated with one unit of BcII. After 30 min, the reaction
afforded the expected hydrolyzed product, as confirmed by 1H NMR
spectroscopy and HPLC experiments. The reactions of oxacillin with the
serine-containing b-lactamase, and of penicillin with both enzymes were
performed by following the same procedure. The hydrolyzed products
were purified by HPLC by using a semipreparative column and the sol-
vent was evaporated under high vacuum. Oxacillin turnover product:
1H NMR (400 MHz, CD3OD): d=1.31 (s, 3H), 1.60 (s, 3H), 2.69 (s, 3H),
3.60 (s, 1H), 4.54 (d, 1H), 5.20 (d, 1H), 7.48–7.49 (m, 3H), 7.86–
7.87 ppm (m, 2H).
Determination of b-lactamase activity: The kinetics of hydrolysis of oxa-
cillin and penicillin G were monitored by HPLC by using a reverse-phase
column. All reactions with the natural enzymes were carried out in phos-
phate buffer and the reactions with synthetic compounds were carried
out in a 9:1 mixture of 0.05m HEPES buffer/DMSO at pH 7.5. The stock
solutions of test complexes and oxacillin were prepared in the appropri-
ate solvent and were used immediately to avoid any possible decomposi-
tion. The concentration of stock solutions of oxacillin, complexes, and in-
hibitor was fixed at 5”10À3 m. In a typical kinetics experiment, a sample
vial containing 1.5 mL of the test complex and an appropriate concentra-
tion of oxacillin was incubated. At various time intervals, 10 mL aliquots
were removed from the reaction mixture and injected directly onto the
HPLC column. The compounds were eluted in a linear-gradient mode
with a mixture of 30–80% acetonitrile and 0.1% trifluoroacetic acid
(TFA) over 8.5 min at a flow rate of 1.0 mLminÀ1. The reaction products
(oxacillin turnover product, retention time: 3.66 min) and oxacillin (re-
tention time: 4.24 min) were separable. The chromatograms were ob-
tained at 254 nm and the concentration of oxacillin or hydrolyzed prod-
uct was determined for each injection from the peak area by using a cali-
bration plot. To avoid any significant changes in the concentrations of
the starting materials or in the product inhibition, only the first 5–10%
of conversion was monitored in most cases. To calculate the percentage
conversion in the presence of the enzymes and complexes, the test sam-
ples were injected onto the column at regular time intervals and the
HPLC was run for several hours. The maximum velocity (Vmax), Michae-
lis constant (Km), turnover number (kcat), and catalytic efficiency (h) for
each complex were calculated from the double-reciprocal or Linewea-
ver–Burk plots.
Chem. Eur. J. 2006, 12, 7797 – 7806
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7805

G. Mugesh et al.
2001, 123, 9867–9879; c) M. D. Peraro, A. J. Vila, P. Carloni, Inorg.
Chem. 2003, 42, 4245–4247.
[25] S. Bounaga, A. P. Laws, M. Galleni, M. I. Page, Biochem. J. 1998,
331, 703–711.
[26] SMART (Version 5.05), Bruker AXS: Madison, WI, 1998.
[27] a) G. M. Sheldrick, SHELXTL (V97–2), Siemens Industrial Auto-
mation, Madison, WI, 1997; b) G. M. Sheldrick, SHELX-97, Pro-
gram for Crystal Structure Solution and Refinement, University of
Gçttingen, Gçttingen (Germany), 1997.
[28] G. M. Sheldrick, SADABS (Version 2), Multi-Scan Absorption Cor-
rection Program, University of Gçttingen, Gçttingen (Germany),
2001.
[20] a) B. Singh, J. R. Long, F. F. de Biani, D. Gatteschi, P. Stavropoulos,
J. Am. Chem. Soc. 1997, 119, 7030–7047; b) A. C. Rosenzweig, P.
Nordlund, P. M. Takahara, C. A. Frederick, S. J. Lippard, Chem.
Biol. 1995, 2, 409–418.
[21] a) R. B. Davis, E. P. Abraham, J. Melling, Biochem. J. 1974, 143,
129–135; b) E. G. Orellano, J. E. Girardini, J. A. Cricco, E. A. Cec-
carelli, A. J. Vila, Biochemistry 1998, 37, 10173–10180.
[22] S. Bounaga, A. P. Laws, M. Galleni, M. I. Page, Biochem. J. 1998,
331, 61–68.
[29] A. L. Spek, J. Appl. Crystallogr. 2003, 36, 7–13.
[30] C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National
Laboratory, Oak Ridge, TN, 1976.
[23] R. M. Rasia, A. J. Vila, J. Biol. Chem. 2004, 279, 26046–26051.
[24] N. V. Kaminskaia, C. He, S. J. Lippard, Inorg. Chem. 2000, 39, 3365–
3373.
Received: May 4, 2006
Published online: August 11, 2006
7806
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7797 – 7806