N-arylsulfonyl hydrazones as inhibitors of IMP-1 metallo-β-lactamase

Article abstract of DOI:10.1128/AAC.46.8.2450-2457.2002

Members of a family of N-arylsulfonyl hydrazones have been identified as novel inhibitors of IMP-1, a metallo-β-lactamase of increasing prevalence. Structure-activity relationship studies have indicated a requirement for bulky aromatic substituents on each side of the sulfonyl hydrazone backbone for these compounds to serve as efficient inhibitors of IMP-1. Molecular modeling has provided insight into the structural basis for the anti-metallo-β-lactamase activity exhibited by this class of compounds.

Full text of DOI:10.1128/AAC.46.8.2450-2457.2002

ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Aug. 2002, p. 2450–2457
Vol. 46, No. 8
0
066-4804/02/$04.00ϩ0 DOI: 10.1128/AAC.46.8.2450–2457.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
N-Arylsulfonyl Hydrazones as Inhibitors of
IMP-1 Metallo-␤-Lactamase
1
1
1
2
Stefan Siemann, Darryl P. Evanoff, Laura Marrone, Anthony J. Clarke,
1
1
Thammaiah Viswanatha, and Gary I. Dmitrienko *
1
Department of Chemistry, University of Waterloo, Waterloo, Ontario N2L 3G1, and Canadian Bacterial Diseases Network,
2
Department of Microbiology, University of Guelph, Guelph, Ontario N1G 2W1, Canada
Received 28 September 2001/Returned for modification 12 January 2002/Accepted 9 April 2002
Members of a family of N-arylsulfonyl hydrazones have been identified as novel inhibitors of IMP-1, a
metallo-␤-lactamase of increasing prevalence. Structure-activity relationship studies have indicated a require-
ment for bulky aromatic substituents on each side of the sulfonyl hydrazone backbone for these compounds to
serve as efficient inhibitors of IMP-1. Molecular modeling has provided insight into the structural basis for the
anti-metallo-␤-lactamase activity exhibited by this class of compounds.
The ever-increasing incidence of microbial resistance to
-lactam antibiotics has been traced, in a large number of
cies (CphA and ImiS) belong to the B2 subfamily, while the
Stenotrophomonas maltophilia enzyme (L1) serves as a proto-
type of the B3 subfamily. The three-dimensional structures of
several metallo-␤-lactamases, including IMP-1, L1, CcrA, and
BcII, have been elucidated by X-ray crystallographic analysis,
revealing the overall protein folds to be similar (7, 43).
Some of the reported inhibitors of these enzymes include
trifluoromethyl alcohols and ketones (41), amino acid-derived
hydroxamates (42), thiols (3, 4, 12, 13), thioester derivatives
(13, 14, 31, 32), cysteinyl peptides (5), biphenyl tetrazoles (38,
39), mercaptocarboxylates (24, 30), 1␤-methylcarbapenem de-
rivatives (25, 26), a synthetic cephamycin (34), and 2,3-disub-
stituted succinic acid derivatives (40).
With a view to developing potent inhibitors of metallo-␤-
lactamases, we initiated the design, synthesis, and systematic
screening of N-arylsulfonyl hydrazones (hereafter referred to
as sulfonyl hydrazones) for their ability to inhibit IMP-1. The
choice of this particular enzyme as the target was motivated by
its impending emergence in other pathogenic organisms via the
facile interspecies transfer of the plasmid-borne blaIMP gene
(19, 44). This report concerns a detailed analysis of the struc-
ture-activity relationship of sulfonyl hydrazones as inhibitors of
IMP-1. Molecular modeling studies, based on the recently pub-
lished crystal structure of an IMP-1 enzyme–mercaptocarboxy-
late inhibitor complex (7), were used to correlate the observed
structure-activity relationship with interactions between these
compounds and the enzyme.
␤
instances, to the ability of the organisms to produce ␤-lacta-
mases, enzymes capable of deactivating these antibiotics by
hydrolysis (16, 22, 28). Consequently, there is a considerable
interest in the development of ␤-lactamase inhibitors that
would allow for the continued use of ␤-lactam antibiotics in
clinical therapy. To date, four distinct classes, designated A, B,
C, and D, of ␤-lactamases have been identified (1, 22, 28). The
members of the A, C, and D classes have been shown to use a
serine residue as a nucleophile in their catalytic mechanism. In
sharp contrast, class B ␤-lactamases are metalloenzymes, with
zinc ions fulfilling their metal requirement. Studies have been
directed for the most part toward screening and/or develop-
ment of potential inhibitors of class A ␤-lactamases. As a
consequence of these studies, clavulanate and penicillanic acid
sulfones (e.g., sulbactam and tazobactam), which function as
potent mechanism-based inhibitors of these enzymes (6, 16),
have been successfully exploited in clinical medicine for the
continued use of penicillins in combating bacterial infections
(
22, 30).
During the past few years, class B ␤-lactamases have at-
tracted considerable attention in view of their ability to func-
tion as carbapenemases (9, 35) as well as their ability to utilize
inhibitors of class A enzymes as substrates (33). All metallo-
␤
-lactamases isolated to date are chromosome encoded, with
the exception of the IMP and VIM enzymes found in certain
strains of Pseudomonas aeruginosa, Serratia marcescens, and
other gram-negative bacteria, which are of plasmid origin (21).
Based on the differences in their primary structures, these
enzymes have been classified into three subfamilies (11). The
B1 subfamily comprises some of the most extensively investi-
gated metallo-␤-lactamases (the commonly used abbreviated
names are given in parentheses), such as those derived from
Bacteroides fragilis (CcrA), Bacillus cereus (BcII), and P. aerugi-
nosa (IMP-1). The metallo-␤-lactamases from Aeromonas spe-
MATERIALS AND METHODS
Materials. Nitrocefin was obtained from Oxoid (Basingstoke, United King-
dom). Glutathione S-transferase gene fusion vector pGEX-4T-3, glutathione-
Sepharose 4B, and MonoQ ion-exchange resin were purchased from Amersham
Pharmacia (Uppsala, Sweden). pCIP4 was kindly provided by M. Galleni (Labo-
ratoire d’Enzymologie and Centre d’Ing e´ nierie des Prot ´e ines, Institut de Chimie,
Universit ´e de Li `e ge, Li `e ge, Belgium).
␤
-Lactamase preparations. PCR was used to retrieve the gene encoding the B.
cereus 5/B/6 metallo-␤-lactamase (BcII) (20), which is homologous (only 17
amino acid residues being different out of a total of 227 such residues) to the
processed form of the protein from strain 569/H/9 (2, 15). An EcoRI-XhoI
fragment of this PCR product was inserted into pGEX-4T-3 for expression of the
enzyme as a glutathione S-transferase fusion protein (37). The expression of the
fusion protein was induced in Escherichia coli DH5␣, and the protein was re-
covered from the cell extract by adsorption to the affinity matrix, glutathione-
*
Corresponding author. Mailing address: Department of Chemis-
try, University of Waterloo, 200 University Ave. W., Waterloo, Ontario
N2L 3G1, Canada. Phone: (519) 888-4642. Fax: (519) 746-0435. E-
mail: dmitrien@muskie.uwaterloo.ca.
2
450

VOL. 46, 2002
METALLO-␤-LACTAMASE INHIBITION
2451
Sepharose 4B. Following the recovery of metallo-␤-lactamase by treatment with
thrombin (37), the protein was further purified by chromatography on an anion-
exchange matrix, MonoQ.
TABLE 1. Influence of sulfonyl hydrazones on IMP-1-mediated
a
nitrocefin hydrolysis
IMP-1 metallo-␤-lactamase was expressed in E. coli BL21(DE3) carrying
pCIP4 and purified as reported by Laraki et al. (18).
The homogeneity of metallo-␤-lactamases was confirmed by sodium dodecyl
sulfate-polyacrylamide gel electrophoresis according to the method of Laemmli
(17). Electrospray ionization (ESI) mass spectrometry (MS) was used to verify
the masses of the purified proteins: (i) IMP-1, 25,112 Ϯ 3 Da, in agreement with
that expected from the amino acid sequence (19); and (ii) BcII, 25,478 Ϯ 2 Da,
in agreement with that expected for BcII ␤-lactamase (20) with a five-amino-
acid-residue segment (Gly-Ser-Pro-Asn-Ser) from vector pGEX-4T-3 at the N
Compound
R1
IC50 (␮M)
terminus of the protein. The kinetic parameters for nitrocefin hydrolysis (kcat, 41
Ϫ1
1
Ph (4-CH3)
40 Ϯ 3.5
16 Ϯ 1.3
48 Ϯ 4.6
40 Ϯ 3.8
25 Ϯ 2.4
s
m
; K , 49 ␮M) obtained with the BcII enzyme are similar to those reported for
2
3
4
5
a
Ph [4-C(CH3)3]
the wild-type enzyme preparation (29).
Ph (4-NO2)
Ph (4-OCH3)
Ph (4-OCF3)
Enzyme concentrations were determined from the absorbance at 280 nm (ε280
values of 44,380 MϪ1 cm and 30,440 M cm for IMP-1 and BcII, respec-
Ϫ1
Ϫ1
Ϫ1
tively).
Synthesis of inhibitors. (i) General procedure for the preparation of arylsul-
fonyl hydrazides. To a stirred solution of anhydrous hydrazine (12.5 mmol) in
CH
2 2
Cl (5 ml) was added a solution of the appropriate sulfonyl chloride (2.5
mmol) over 2 min. The reaction mixture was stirred for 15 min, and the pH was
ESI MS. ESI MS studies were performed with a QuattroII mass spectrometer
Micromass, Manchester, United Kingdom) equipped with an ESI source.
Molecular modeling. The IMP-1–inhibitor (sulfonyl hydrazone) structural
subsequently adjusted to approximately 11 with 10% Na
separated, and the aqueous phase was extracted with CH
2
CO
Cl
3
. The layers were
(three times with
(
2
2
2
4
5 ml each time). The combined organic layers were dried over MgSO and
model was based on the recently reported X-ray crystal structure of the IMP-1
enzyme from P. aeruginosa in complex with a mercaptocarboxylate inhibitor (7).
To construct a model of the inhibitor, a single X-ray crystal diffraction study of
sulfonyl hydrazone 1 (Table 1) was performed. The obtained coordinates were
used to generate a model of the hypothetical bound state by manually docking
the hydrazone structure into the protein crystal structure containing the bound
mercaptocarboxylate inhibitor (entry code, 1DD6) such that one of the sulfonyl
filtered, and the solvent was removed under reduced pressure. No further puri-
fication of the product was required.
(ii) General procedure for the preparation of arylsulfonyl hydrazones. A
mixture of the required sulfonyl hydrazide (0.5 mmol) and the appropriate
aldehyde (0.5 mmol) was dissolved in hot ethyl alcohol (2 ml). After the mixture
was cooled to room temperature, the product was precipitated by the dropwise
addition of water. Following filtration, the precipitate was washed with cold H
2 ml) and hexane (2 ml) and dried in vacuo.
All synthetic compounds were fully characterized by 1H nuclear magnetic
2
O
group oxygen atoms (Pro-R) was positioned to establish contacts with both zinc
ions (OOZn1, 2.278 A˚ ; OOZn2, 2.260 A˚ ), similar to those formed by the thiol
(
1
3
group of the mercaptocarboxylate inhibitor (7). The mercaptocarboxylate struc-
ture was subsequently extracted, hydrogen atoms were added to all appropriate
sites, and minimization of the resulting hydrazone-protein complex was carried
out by using the MMFF94 molecular mechanics force field and the Anneal
function within Sybyl 6.7 (Tripos, Inc., St. Louis, Mo.). Additional models of
bound sulfonyl hydrazones were constructed by modifying the inhibitor structure
within the protein-inhibitor complex, followed by minimization, as described
above.
resonance (NMR) spectrometry (300 or 500 MHz), C NMR spectrometry (75
MHz), and MS. Mass spectra were recorded at the McMaster University Mass
Spectrometry Laboratory, Hamilton, Ontario, Canada.
Metallo-␤-lactamase assays. A spectrophotometric method based on the use
of a chromophoric substrate, nitrocefin (27), was used for the measurement of
␤
-lactamase activity with the aid of a Cary5 spectrophotometer (Varian, Missis-
sauga, Ontario, Canada). All assays were performed with HEPES buffer (50 mM,
pH 7.3) at 30°C. The procedure involved the preparation of the following stock
solutions: (i) nitrocefin (2 mM) in HEPES buffer (50 mM, pH 7.3) containing
dimethyl sulfoxide (DMSO) (5% [vol/vol]) and (ii) metallo-␤-lactamase (400
Amino acid residues are labeled according to the recently proposed class B
␤
-lactamase numbering scheme (11). For convenience, the formerly used num-
bering of IMP-1 amino acid residues is given in parentheses.
nM) in HEPES buffer containing ZnCl
2
(100 ␮M) and bovine serum albumin
(BSA) (100 ␮g/ml). The inclusion of BSA in the enzyme stock solution results in
the protection of ␤-lactamase from denaturation at low protein concentrations
and thus allows for a reliable and reproducible assessment of the initial rate of
enzyme-catalyzed hydrolysis of its substrate, observations in agreement with
those reported by Laraki et al. (18). At the low concentration (1 ␮g/ml) used in
the assay, BSA showed no detectable interaction with the product of nitrocefin
hydrolysis. Furthermore, the degree of inhibition was not significantly influenced
by the presence of BSA up to a concentration, in the assay, of at least 20 ␮g/ml.
In a typical assay (final volume, 1 ml), metallo-␤-lactamase (4 nM) in HEPES
buffer was allowed to equilibrate at 30°C for 1 min. The reaction was initiated by
the introduction of the substrate (100 ␮M for IMP-1 and 200 ␮M for BcII), and
the increase in the absorbance at 482 nm was recorded.
For determination of the concentration required to effect 50% inhibition of
enzyme activity (IC50), the enzyme was preincubated with the desired compound
for 1 min at 30°C prior to the initiation of the assay by the addition of the
substrate. The compounds to be tested were dissolved in DMSO prior to incu-
bation with the enzyme, with the final concentration of DMSO in the assay not
exceeding 1% (vol/vol). IC50s were deduced from a plot of percent loss of activity
versus inhibitor concentration. The precision of these determinations was within
the range of Ϯ10%.
RESULTS AND DISCUSSION
Optimization of assay conditions. The effect of any putative
inhibitor on a target enzyme is generally assessed in reference
to control experiments performed in its absence. Since the
determination of the inhibitory potency of a given compound
involves its preincubation with the target enzyme, the effect of
such preexposure on the catalytic efficiency of IMP-1 appeared
essential, especially in view of its lability at very low concen-
trations.
Such studies have revealed that the kinetic parameters (Km
and kcat) depend on whether the assay is initiated by the ad-
dition of the enzyme to the substrate (no preincubation) or
started by the addition of the substrate (prewarmed to 30°C) to
the enzyme maintained at 30°C for 1 min (preincubation). In
the former situation, K and kcat values were 21 Ϯ 2 ␮M and
Determination of steady-state kinetic constants. Estimates of steady-state
m
Ϫ1
kinetic parameters for the hydrolysis of nitrocefin (K
m
and kcat) were achieved by
320 Ϯ 30 s , respectively, while under the latter conditions,
these values were found to be 12 Ϯ 1 ␮M and 260 Ϯ 22 s
respectively. The latter values are in excellent agreement with
those recorded in a recent report (23). Extension of the period
of preincubation of IMP-1 beyond 1 min did not lead to further
changes in the kinetic parameters of the enzyme. Hence, all
Ϫ1
fitting the initial velocity data to the Michaelis-Menten equation with the soft-
ware package Grafit 4.0 (Erithacus Software Ltd., Staines, United Kingdom).
Inhibitor constants (K ) were assessed according to the procedure of Dixon (10).
i
,
The resulting data were used to generate a Cornish-Bowden plot (8) to establish
Ϫ1
the mode of inhibition exhibited by the compounds. A ⌬ε482 value of 15,930 M
cm for nitrocefin hydrolysis was used in the estimation of kcat values.
Ϫ1

2
452
SIEMANN ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
TABLE 2. Effect of replacement of a 2-hydroxynaphthyl group on
the inhibitory activity of sulfonyl hydrazones against IMP-1
Compound
R
2
IC50 (␮M)
17.5 Ϯ 1.6
6
a
FIG. 1. Structural representation of sulfonyl hydrazone derivatives
(
a) and cephalosporins (b). Bonds and atoms in bold depict structural
similarities between the classes of compounds.
6
b
d
6.3 Ϯ 0.6
10.5 Ϯ 1.0
Ͼ250
6
experiments with IMP-1 were routinely performed subsequent
to its preincubation for 1 min at 30°C.
In experiments which involved the testing of sulfonyl hydra-
zones for their inhibitory activity, the enzyme was preincubated
with the desired compound for 1 min at 30°C (conditions
similar to those noted above) prior to initiation of the assay by
the addition of the substrate. Extending the period of incuba-
tion of the enzyme with the inhibitor beyond 1 min did not lead
to further changes in enzymatic activity, indicating that sulfonyl
hydrazones are relatively fast in exerting their inhibitory effect.
Structure-inhibitory activity relationship. Two aspects of
sulfonyl hydrazones, (i) the resemblance of their backbone
structural elements to those in the scissile segment of cepha-
losporins, the natural substrates of these enzymes (Fig. 1), and
6e
were moderately affected by the nature of the substituent lo-
cated on the aryl moiety of segment A. Compounds bearing
larger substituents, such as a t-butyl group (compound 2a) or
an OCF group (compound 5), tended to be more potent than
3
other inhibitors listed in Table 1. It is interesting that the
propensity of the para substituent to donate or withdraw elec-
trons appears to play a less important role in tuning the inhib-
itory potencies of the compounds tested. On the other hand,
replacement of the 2-hydroxynaphthyl substituent in segment
B (as in compound 6a) with a 2-hydroxybenzene ring (as in
compound 6e) resulted in more than a 15-fold increase in the
IC50 (Table 2), suggesting the need for a bulky substituent at
this location for the compound to be an effective inhibitor.
In order to further assess the contribution of the 2-hydroxy
group in segment B to the anti-metallo-␤-lactamase activity,
studies were extended to two analogs of compound 6a, one in
which the hydroxyl substituent was deleted (compound 6d) and
the other in which it was masked by methylation (compound
6b). Interestingly, both compounds, 6d and 6b, were slightly
more effective as inhibitors of IMP-1 than the one with the
hydroxyl group, 6a (Table 2). These findings tend to negate the
earlier noted premise of the hydroxyl group at this location
serving as an equivalent for the carboxyl function of cephalo-
sporin substrates. In addition, its involvement as a metal bind-
ing ligand appears remote, especially since the enzymatic ac-
tivity (with or without an inhibitor) of the protein is not
significantly affected by the inclusion of 0.1 mM extraneous
(ii) their potential to function as moderate binding agents for
divalent metal ions, provided an impetus for screening these
compounds for their ability to function as novel inhibitors of
metallo-␤-lactamases, with the IMP-1 enzyme serving as the
target protein.
For N-arylsulfonyl amino acids, metal complexation at high
pH has been shown to involve the deprotonated sulfonamide
nitrogen and the carboxylate function of the amino acid moiety
(reference 36 and references therein). With regard to the se-
lectivity in their action, arylsulfonyl compounds normally ex-
hibit a weak affinity for zinc ions (36), a feature that would
prevent them from interfering in the functioning of other Zn-
containing proteins. However, their propensity to acquire an
enhanced affinity for metal ions when present in a proper
orientation could render them target specific in their function.
The sulfonyl hydrazones initially chosen for study of their
potential anti-metallo-␤-lactamase activity are listed in Table
1
. For clarity in presentation, the sulfonyl hydrazine moiety of
these compounds is designated segment A, while the compo-
nent which contributes the carbonyl function for condensation
with the hydrazide is designated segment B (Fig. 1a). The
ligands for metal binding may be provided by either one or
both of these segments. In segment A, the deprotonated sul-
fonamide and/or the sulfonyl oxygens might fulfill this func-
tion. The hydroxyl group in segment B might serve either as a
metal binding ligand or as a crude mimic of the carboxylate
group of cephalosporins (Fig. 1).
2
ϩ
Zn (data not shown). Furthermore, screening of some of the
sulfonyl hydrazones examined in this study for their ability to
2
ϩ
inhibit the mammalian Zn -dependent enzyme carboxypep-
tidase A revealed these compounds to possess such low inhib-
itory activity that IC s could not be obtained in most instances
5
0
The details concerning the structure-activity relationship of
the initially investigated sulfonyl hydrazones with substitutions
on the benzene ring with IMP-1 are provided in Table 1. As
indicated by the IC s, the potencies of sulfonyl hydrazones
due to precipitation of the inhibitors at high concentrations.
When IC s could be estimated (e.g., with compound 11c),
they were Ͼ10-fold those noted with IMP-1. Thus, these ob-
5
0
servations are consistent with the assertion that the inhibition
5
0

VOL. 46, 2002
METALLO-␤-LACTAMASE INHIBITION
2453
TABLE 3. Inhibitory activity of various 2-hydroxynaphthyl, 2 methoxynaphthyl, and anthracyl derivatives on
IMP-1-mediated nitrocefin hydrolysis
a
The precision of the IC50 determinations was within the range of Ϯ10%.
of IMP-1 by sulfonyl hydrazones is not the result of nonspecific
metal ion binding.
fonyl hydrazones, with the anti-metallo-␤-lactamase activity
being directly proportional to the size of the substituent and
inversely proportional to its electron-withdrawing property.
Furthermore, the 2,5-dichloro derivatives 10, which tended to
be slightly more potent than the para-chloro derivatives 8,
exhibited potencies similar to those observed for the 4-bromo
derivatives 7.
The most potent inhibitors within the series of compounds
listed in Table 3 were 11c and 12c, each with an IC50 of 1.6 ␮M.
These results are consistent with the observation that large
substituents with less electron-withdrawing capacity improve
the inhibition of the IMP-1 metallo-␤-lactamase by sulfonyl
hydrazone compounds. Sulfonyl hydrazones with a 2-substi-
tuted naphthyl moiety in segment A appear to be more effec-
tive as inhibitors than their 1-substituted counterparts (com-
pare compounds 12 and 13). In addition, the introduction of a
bulky dimethylamino substituent at position 5 of the 1-substi-
tuted naphthyl ring system, as in compound 14 (Fig. 2), led to
a significant diminution of inhibitory potency. Such an adverse
effect was also noted for the electron-withdrawing 8-quinolyl
derivative (compound 15) (Fig. 2). In summary, these findings
In an effort to further elucidate the structural features nec-
essary for improved inhibition and thus to develop more potent
metallo-␤-lactamase inhibitors, a series of each of the 2-hy-
droxynaphthyl, 2-methoxynaphthyl, and anthracyl sulfonyl hy-
drazone derivatives with a variety of substitutions in the aro-
matic moiety in segment A were prepared, and their potential
to inhibit IMP-1 was determined (Table 3). The rationale for
the replacement of the naphthyl substituent with an anthracyl
substituent was to elicit the influence of the hydrophobicity
and/or the size of the aromatic ring system on the inhibitory
potency of the compounds. As shown in Table 3, all anthracyl
derivatives exhibited higher inhibitory activity (the IC s de-
5
0
creased approximately two- to fourfold) than the correspond-
ing 2-methoxynaphthyl derivatives (b versus c). Replacement
of the para-iodo substituent with the lighter halogen atoms
resulted in a substantial decrease in inhibitory efficiency, in the
following order: I Ն Br Ͼ Cl Ͼ F (compounds 6 to 9). This
observation reinforces the view that the nature of the para
substituents plays an important role in the binding of the sul-

2
454
SIEMANN ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
poor inhibitor of BcII, causing approximately 30% inhibition at
a concentration of 25 ␮M; other sulfonyl hydrazones (e.g.,
compounds 5, 6a or 6d, 9a, 11a, and 12a) were marginal in their
action, with IC s of Ն100 ␮M. In this connection, it is inter-
5
0
esting that BcII (from strain 5/B/6), despite the presence of two
Zn(II) binding sites, has been shown to be catalytically fully
functional with just one metal ion bound to the active site (29).
Although the catalytic competence of a mono-Zn form of
IMP-1 remains to be established, the kinetic parameters of the
binuclear enzyme have been documented (18). At the present
time, it is not apparent whether the observed difference in the
inhibitory potencies of sulfonyl hydrazones toward IMP-1 and
BcII is a reflection of variations in the geometry of the active
site(s) or is related to a change in the metal requirement for
catalytic function.
Molecular modeling. The recent release of the coordinates
of IMP-1 in complex with a mercaptocarboxylate inhibitor (7)
prompted us to construct a model for a possible mode of
binding of sulfonyl hydrazones to the enzyme. Initially, a hy-
pothetical model of compound 1 bound to the active site of
IMP-1 was constructed as described in Materials and Methods.
As shown in Fig. 4A, the Pro-S sulfonyl oxygen of compound
FIG. 2. Structural representation of compounds 14 and 15 and
their IC s.
5
0
suggest that the inhibitory potency of sulfonyl hydrazones is
sensitive to alterations in the orientation, size, and electronic
properties of the aryl moiety in segment A.
Mode of inhibition of the IMP-1 metallo-␤-lactamase by
sulfonyl hydrazone compounds. Compounds 6a and 11c were
selected to identify the mode of their interactions with IMP-1.
Analysis of the data with Dixon plots (Fig. 3A, compound 6a)
indicated that both compounds function as reversible enzyme
inhibitors, with calculated K values of 6.6 Ϯ 1.1 and 0.7 Ϯ 0.1
i
␮M for compounds 6a and 11c, respectively. The mode of
1
forms an H bond to the amide group of the side chain of
inhibition was determined with the aid of Cornish-Bowden
plots (Fig. 3B, compound 6a), which showed that compounds
Asn233 (167), consistent with the assumption that the sulfon-
amide group in this inhibitor may serve as a mimic of a tetra-
hedral intermediate involved in the hydrolysis of substrates
6
a and 11c serve as mixed-type inhibitors of IMP-1, with KiЈ
(dissociation constant for the enzyme-substrate-inhibitor com-
(28, 43). Additional significant electrostatic stabilization of the
plex) values of 25.3 Ϯ 2.0 and 2.2 Ϯ 0.2 ␮M, respectively. Since
the magnitude of K is considerably smaller than that of K Ј, it
negative charge on the nitrogen atom of the deprotonated
sulfonamide group may arise by interaction with Zn1 and the
i
i
would appear that the observed inhibition of IMP-1 by these
compounds is mainly due to their specific interactions with the
active site of the enzyme.
Studies with B. cereus 5/B/6 metallo-␤-lactamase. The ability
of sulfonyl hydrazones to inhibit the metallo-␤-lactamase from
B. cereus (5/B/6), BcII, was also investigated. These studies
revealed that these compounds had little or no adverse effect
on the activity of BcII. Thus, compound 11c, which was most
protonated ε-NH group of Lys224 (161).
2
Examination of the naphthyl ring of the inhibitor in this
complex reveals a van der Waals (VdW) contact between C -H
4
of the naphthyl ring and the indole ring nitrogen of Trp64 (28),
a residue which is part of a connecting turn (flap) linking two
antiparallel ␤ strands involved in inhibitor-substrate binding. A
very similar interaction can be expected to occur in the com-
plex of the salicylaldehyde-derived analog, compound 6e.
effective in inhibiting IMP-1 (K Ϸ 0.7 ␮M), was found to be a
i
FIG. 3. Dixon (A) and Cornish-Bowden (B) plots illustrating mixed-type inhibition of IMP-1-mediated nitrocefin hydrolysis by compound 6a.
The inhibitor constant (K ) was estimated by graphic analysis of the Dixon plot. Since Dixon plots are not suited for distinguishing between
i
competitive and mixed-type inhibition kinetics, Cornish-Bowden plots were used to determine the mode of inhibition. V, initial rate (velocity) of
the reaction; S, substrate concentration.

VOL. 46, 2002
METALLO-␤-LACTAMASE INHIBITION
2455
FIG. 4. Stereoviews depicting the active sites of the modeled complexes of sulfonyl hydrazones 1 (A) and 2c (B) with the IMP-1 metallo-␤-
lactamase. The models are based on the published coordinates of an enzyme-mercaptocarboxylate inhibitor complex (7). Inhibitors and selected
amino acid residues lining the binding site and discussed in the text are shown in capped-stick representation (carbon, gray; nitrogen, blue; oxygen,
red; sulfur, yellow). The numbers shown denote the positions of amino acid residues in the primary structure of the enzyme. The zinc ions are
depicted as large magenta spheres. Water molecules, which are proposed to contribute to a network of favorable interactions between the inhibitor
and the protein (see the text), are shown as small red spheres.
However, since compound 6e is a poor inhibitor, such an in-
teraction does not appear to be the sole contributor to the
favorable interaction of the naphthyl ring with the active site of
the enzyme. Indeed, a closer examination of the model de-
picted in Fig. 4A reveals that the second benzenoid ring estab-
lishes favorable VdW contacts with both the protein [C-6 of
the naphthalene ring and Gly228 (164)] and two water mole-
cules with C-7 of the naphthalene ring. The latter two mole-
cules are part of a remarkable cluster of five water molecules
filling the space between two loop regions that incorporate
amino acid residues [Cys221 (158), Lys224 (161), His263 (197)]
important for zinc ion and substrate binding. This water cluster
involves H bonds among the water molecules as well as to the
protein and may play an important role in stabilizing the ac-
tive-site region.
of the hydrazone linkage is observed in the model (Fig. 4A;
naphthyl-CAN torsion, 52°), whereas a nearly coplanar ar-
rangement of these groups, stabilized by an intramolecular H
bond between the OH group and the CAN group, is found in
the crystal structure of the free inhibitor. Thus, the comple-
mentary interaction of the inhibitor with the enzyme requires
the concomitant destabilization of the above-mentioned in-
tramolecular H bond. This feature may offer an explanation as
to why the potency of the inhibitor is increased by replacement
of the hydroxynaphthyl ring with a methoxynaphthyl or 9-an-
thracene ring system (Fig. 4B). In the latter two cases, the
planes of the aryl group and the CAN group of the free
inhibitors are intrinsically skewed, eliminating the need to de-
viate significantly from their minimum-energy conformations
in order to bind to the active site of the enzyme.
In this model, the hydroxyl group of the 2-hydroxynaphthyl
system is not near Lys224 (161), a feature that minimizes the
importance of this functional group for the interaction of sul-
fonyl hydrazones with the enzyme. This finding is in full accor-
dance with the structure-activity profile observed for this class
of compounds.
The para substituent on the aromatic ring in this model is
oriented toward a hydrophobic pocket lined by Phe87 (51), the
␤- and ␥-CH groups of Glu59 (23), and one of the ␤-methyl
2
groups of Val61 (25). The higher potencies of inhibitors with
larger para substituents, such as the t-butyl group in compound
2c (Fig. 4B), are explicable largely on the basis of more exten-
sive VdW contacts with this hydrophobic pocket.
It is interesting that a substantial skewing of the plane of the
hydroxynaphthyl ring relative to the plane of the CAN group
In addition to the interactions noted above, the aromatic

2
456
SIEMANN ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
ring attached to the sulfonyl group (segment A) establishes a
favorable contact with the carboxylate carbon atom of the zinc
binding side chain of Asp120 (81). Furthermore, an interesting
interaction is present between this ring and a water molecule
which acts as an H-bond donor to the carboxylate group of
Asp120 (81). This water molecule is oriented so as to serve as
an H-bond donor to the face of the aromatic ring of the aryl
sulfonamide portion of the inhibitor. This specific interaction
may also partly explain the diminution in inhibitory potency
observed with compounds of the para-halogen series in the
order I Ն Br Ͼ Cl Ͼ F, since the electron density in the ring
system and, hence, the H-bond acceptor strength would be
increasingly diminished in inhibitors of the same series.
In summary, the models presented in this study, although
hypothetical, do provide some insights into possible specific
interactions between sulfonyl hydrazones and the enzyme
which help rationalize the observed structure-activity relation-
ship at a molecular level. These insights may provide the basis
for the development of more potent inhibitors of IMP-1.
of thioester and thiol inhibitors of IMP-1 metallo-␤-lactamase. Bioorg. Med.
Chem. Lett. 9:2549–2554.
4. Hammond, G. G., J. L. Huber, M. L. Greenlee, J. B. Laub, K. Young, L. L.
Silver, J. M. Balkovec, K. D. Pryor, J. K. Wu, B. Leiting, D. L. Pompliano,
and J. H. Toney. 1999. Inhibition of IMP-1 metallo-␤-lactamase and sensi-
tization of IMP-1-producing bacteria by thioester derivatives. FEMS Micro-
biol. Lett. 179:289–296.
15. Hussain, M., A. Carlino, M. J. Madonna, and J. O. Lampen. 1985. Cloning
and sequencing of the metallothioprotein ␤-lactamase II gene of Bacillus
cereus 569/H in Escherichia coli. J. Bacteriol. 164:223–229.
6. Knowles, J. R. 1985. Penicillin resistance: the chemistry of ␤-lactamase
inhibition. Acc. Chem. Res. 18:97–104.
7. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of
the head of bacteriophage T4. Nature 227:680–685.
18. Laraki, N., N. Franceschini, G. M. Rossolini, P. Santucci, C. Meunier, E. de
Pauw, G. Amicosante, J. M. Fr `e re, and M. Galleni. 1999. Biochemical char-
acterization of the Pseudomonas aeruginosa 101/1477 metallo-␤-lactamase
IMP-1 produced by Escherichia coli. Antimicrob. Agents Chemother. 43:
902–906.
9. Laraki, N., M. Galleni, I. Thamm, M. L. Riccio, G. Amicosante, J. M. Fr `e re,
and G. M. Rossolini. 1999. Structure of In31, a blaIMP-containing Pseudo-
monas aeruginosa integron phyletically related to In5, which carries an un-
usual array of gene cassettes. Antimicrob. Agents Chemother. 43:890–901.
0. Lim, H. M., J. J. Pene, and R. W. Shaw. 1988. Cloning, nucleotide sequence,
and expression of the Bacillus cereus 5/B/6 ␤-lactamase II structural gene. J.
Bacteriol. 170:2873–2878.
1
1
1
1
2
2
1. Livermore, D. M., and N. Woodford. 2000. Carbapenemases: a problem in
waiting? Curr. Opin. Microbiol. 3:489–495.
ACKNOWLEDGMENTS
2
2. Matagne, A., A. Dubus, M. Galleni, and J.-M. Fr `e re. 1999. The ␤-lactamase
cycle: a tale of selective pressure and bacterial ingenuity. Nat. Prod. Rep.
This work was supported by the Natural Sciences and Engineering
Research Council of Canada.
1
6:1–19.
2
3. Materon, I. C., and T. Palzkill. 2001. Identification of residues critical for
metallo-␤-lactamase function by codon randomization and selection. Protein
Sci. 10:2556–2565.
We are indebted to Gilles Lajoie, Dyanne Brewer, and Amanda
Doherty-Kirby of the Department of Chemistry, University of Water-
loo (currently at the University of Western Ontario, London, Ontario,
Canada), for ESI MS measurements and Souzan Armstrong for pre-
liminary studies concerning the stability of metallo-␤-lactamases under
assay conditions. We also thank Moreno Galleni for providing the
plasmid used in this study as well as Matthew D. R. Brown and
Nicholas J. Taylor for crystallographic data on tosylhydrazone.
24. Mollard, C., C. Moali, C. Papamicael, C. Damblon, S. Vessilier, G. Amico-
sante, C. J. Schofield, M. Galleni, J. M. Fr `e re, and G. C. Roberts. 2001.
Thiomandelic acid, a broad spectrum inhibitor of zinc ␤-lactamases: kinetic
and spectroscopic studies. J. Biol. Chem. 276:45015–45023.
5. Nagano, R., Y. Adachi, H. Imamura, K. Yamada, T. Hashizume, and H.
Morishima. 1999. Carbapenem derivatives as potential inhibitors of various
2
␤
-lactamases, including class B metallo-␤-lactamases. Antimicrob. Agents
Chemother. 43:2497–2503.
REFERENCES
2
6. Nagano, R., Y. Adachi, T. Hashizume, and H. Morishima. 2000. In vitro
antibacterial activity and mechanism of action of J-111,225, a novel 1␤-
methylcarbapenem, against transferable IMP-1 metallo-␤-lactamase produc-
ers. J. Antimicrob. Chemother. 45:271–276.
27. O’Callaghan, C. H., A. Morris, S. M. Kirby, and A. H. Shingler. 1972. Novel
method for detection of ␤-lactamases by using a chromogenic cephalosporin
substrate. Antimicrob. Agents Chemother. 1:283–288.
1
2
. Ambler, R. P. 1980. The structure of ␤-lactamases. Philos. Trans. R. Soc.
London B Biol. Sci. 289:321–331.
. Ambler, R. P., M. Daniel, J. Fleming, J.-M. Hermoso, C. Pang, and S. G.
Waley. 1985. The amino acid sequence of the zinc-requiring ␤-lactamase II
from the bacterium Bacillus cereus 569. FEBS Lett. 189:207–211.
. Arakawa, Y., N. Shibata, K. Shibayama, H. Kurokawa, T. Yagi, H. Fujiwara,
and M. Goto. 2000. Convenient test for screening metallo-␤-lactamase-pro-
ducing gram-negative bacteria by using thiol compounds. J. Clin. Microbiol.
3
28. Page, M. I., and A. P. Laws. 1998. The mechanism of catalysis and the
inhibition of ␤-lactamases. Chem. Commun. 1609–1617.
3
8:40–43.
29. Paul-Soto, R., R. Bauer, J. M. Fr `e re, M. Galleni, W. Meyer-Klaucke, H.
Nolting, G. M. Rossolini, D. de Seny, M. Hernandez-Valladares, M. Zeppe-
4
5
6
7
. Bounaga, S., A. P. Laws, M. Galleni, and M. I. Page. 1998. The mechanism
of catalysis and the inhibition of the Bacillus cereus zinc-dependent ␤-lacta-
mase. Biochem. J. 331:703–711.
2
ϩ
zauer, and H.-W. Adolph. 1999. Mono- and binuclear Zn -␤-lactamase.
Role of the conserved cysteine in the catalytic mechanism. J. Biol. Chem.
274:13242–13249.
. Bounaga, S., M. Galleni, A. P. Laws, and M. I. Page. 2001. Cysteinyl peptide
inhibitors of Bacillus cereus zinc ␤-lactamase. Bioorg. Med. Chem. 9:503–
30. Payne, D. J., W. Du, and J. H. Bateson. 2000. ␤-Lactamase epidemiology and
the utility of established and novel ␤-lactamase inhibitors. Exp. Opin. Inves-
tig. Drugs 9:247–261.
31. Payne, D. J., J. H. Bateson, B. C. Gasson, T. Khushi, D. Proctor, S. C.
Pearson, and R. Reid. 1997. Inhibition of metallo-␤-lactamase by a series of
thiol ester derivatives of mercaptophenylacetic acid. FEMS Microbiol. Lett.
157:171–175.
32. Payne, D. J., J. H. Bateson, B. C. Gasson, D. Proctor, T. Khushi, T. H.
Farmer, D. A. Tolson, D. Bell, P. W. Skett, A. C. Marshall, R. Reid, L.
Ghosez, Y. Combret, and J. Marchand-Brynaert. 1997. Inhibition of metallo-
␤-lactamase by a series of mercaptoacetic acid thiol ester derivatives. Anti-
microb. Agents Chemother. 41:135–140.
33. Prosperi-Meys, C., G. Llabres, D. de Seny, R. Paul-Soto, M. Hernandez-
Valladares, N. Laraki, J. M. Fr `e re, and M. Galleni. 1999. Interaction be-
tween class B ␤-lactamases and suicide substrates of active-site serine ␤-lac-
tamases. FEBS Lett. 443:109–111.
5
10.
. Bush, K., C. Macalintal, B. A. Rasmussen, V. J. Lee, and Y. Yang. 1993.
Kinetic interaction of tazobactam with ␤-lactamases from all major struc-
tural classes. Antimicrob. Agents Chemother. 37:851–858.
. Concha, N. O., C. A. Janson, P. Rowling, S. Pearson, C. A. Cheever, B. P.
Clarke, C. Lewis, M. Galleni, J. M. Fr `e re, D. J. Payne, J. H. Bateson, and
S. S. Abdel-Meguid. 2000. Crystal structure of the IMP-1 metallo ␤-lacta-
mase from Pseudomonas aeruginosa and its complex with a mercaptocar-
boxylate inhibitor: binding determinants of a potent, broad-spectrum inhib-
itor. Biochemistry 39:4288–4298.
8
9
. Cornish-Bowden, A. 1974. A simple graphical method for determining the
inhibition constants of mixed, uncompetitive and non-competitive inhibitors.
Biochem. J. 137:143–144.
. Cricco, J. A., and A. J. Vila. 1999. Class B ␤-lactamases: the importance of
being metallic. Curr. Pharm. Des. 5:915–927.
1
1
0. Dixon, M. 1953. The determination of enzyme inhibitor constants. Biochem.
J. 55:170–171.
1. Galleni, M., J. Lamotte-Brasseur, G. M. Rossolini, J. Spencer, O. Dideberg,
J. M. Fr `e re, et al. 2001. Standard numbering scheme for class B ␤-lactama-
ses. Antimicrob. Agents Chemother. 45:660–663.
34. Quiroga, M. I., N. Franceschini, G. M. Rossolini, G. Gutkind, G. Bonfiglio,
L. Franchino, and G. Amicosante. 2000. Interactions of cefotetan and the
metallo-␤-lactamases produced in Aeromonas spp. and in vitro activity. Che-
motherapy 46:177–183.
35. Rasmussen, B. A., and K. Bush. 1997. Carbapenem-hydrolyzing ␤-lactama-
ses. Antimicrob. Agents Chemother. 41:223–232.
36. Saladini, M., D. Iacopino, and L. Menabue. 2000. Metal(II) binding ability of
a novel N-protected amino acid. A solution-state investigation on binary and
ternary complexes with 2,2Ј-bipyridine. J. Inorg. Biochem. 78:355–361.
37. Smith, D. B., and K. S. Johnson. 1988. Single-step purification of polypep-
1
2. Goto, M., T. Takahashi, F. Yamashita, A. Koreeda, H. Mori, M. Ohta, and
Y. Arakawa. 1997. Inhibition of the metallo-␤-lactamase produced from
Serratia marcescens by thiol compounds. Biol. Pharm. Bull. 20:1136–1140.
3. Greenlee, M. L., J. B. Laub, J. M. Balkovec, M. L. Hammond, G. G. Ham-
mond, D. L. Pompliano, and J. H. Epstein-Toney. 1999. Synthesis and SAR
1

VOL. 46, 2002
METALLO-␤-LACTAMASE INHIBITION
metallo-␤-lactamase. J. Biol. Chem. 276:31913–31918.
41. Walter, M. W., A. Felici, M. Galleni, R. Paul-Soto, R. M. Adlington, J. E.
Baldwin, J. M. Fr `e re, M. Gololobov, and C. J. Schofield. 1996. Trifluoro-
methyl alcohol and ketone inhibitors of metallo-␤-lactamases. Bioorg. Med.
Chem. Lett. 6:2455–2458.
42. Walter, M. W., M. Hernandez-Valladares, R. M. Adlington, G. Amicosante,
J. E. Baldwin, J. M. Fr `e re, M. Galleni, G. M. Rossolini, and C. J. Schofield.
1999. Hydroxamate inhibitors of Aeromonas hydrophila AE036 metallo-␤-
lactamase. Bioorg. Chem. 27:35–40.
2457
tides expressed in Escherichia coli as fusions with glutathione S-transferase.
Gene 67:31–40.
3
8. Toney, J. H., P. M. Fitzgerald, N. Grover-Sharma, S. H. Olson, W. J. May,
J. G. Sundelof, D. E. Vanderwall, K. A. Cleary, S. K. Grant, J. K. Wu, J. W.
Kozarich, D. L. Pompliano, and G. G. Hammond. 1998. Antibiotic sensiti-
zation using biphenyl tetrazoles as potent inhibitors of Bacteroides fragilis
metallo-␤-lactamase. Chem. Biol. 5:185–196.
9. Toney, J. H., K. A. Cleary, G. G. Hammond, X. Yuan, W. J. May, S. M.
Hutchins, W. T. Ashton, and D. E. Vanderwall. 1999. Structure-activity
relationships of biphenyl tetrazoles as metallo-␤-lactamase inhibitors.
Bioorg. Med. Chem. Lett. 9:2741–2746.
3
4
43. Wang, Z., W. Fast, A. M. Valentine, and S. J. Benkovic. 1999. Metallo-␤-
lactamase: structure and mechanism. Curr. Opin. Chem. Biol. 3:614–622.
44. Watanabe, M., S. Iyobe, M. Inoue, and S. Mitsuhashi. 1991. Transferable
imipenem resistance in Pseudomonas aeruginosa. Antimicrob. Agents Che-
mother. 35:147–151.
0. Toney, J. H., G. G. Hammond, P. M. D. Fitzgerald, N. Sharma, J. M.
Balkovec, G. P. Rouen, S. H. Olson, M. L. Hammond, M. L. Greenlee, and
Y.-D. Gao. 2001. Succinic acids as potent inhibitors of plasmid-borne IMP-1