4-Amino-1,2,4-triazole-3-thione-derived Schiff bases as metallo-β-lactamase inhibitors

Article abstract of DOI:10.1016/j.ejmech.2020.112720

Resistance to β-lactam antibiotics in Gram-negatives producing metallo-β-lactamases (MBLs) represents a major medical threat and there is an extremely urgent need to develop clinically useful inhibitors. We previously reported the original binding mode of 5-substituted-4-amino/H-1,2,4-triazole-3-thione compounds in the catalytic site of an MBL. Moreover, we showed that, although moderately potent, they represented a promising basis for the development of broad-spectrum MBL inhibitors. Here, we synthesized and characterized a large number of 4-amino-1,2,4-triazole-3-thione-derived Schiff bases. Compared to the previous series, the presence of an aryl moiety at position 4 afforded an average 10-fold increase in potency. Among 90 synthetic compounds, more than half inhibited at least one of the six tested MBLs (L1, VIM-4, VIM-2, NDM-1, IMP-1, CphA) with K_i values in the μM to sub-μM range. Several were broad-spectrum inhibitors, also inhibiting the most clinically relevant VIM-2 and NDM-1. Active compounds generally contained halogenated, bicyclic aryl or phenolic moieties at position 5, and one substituent among o-benzoic, 2,4-dihydroxyphenyl, p-benzyloxyphenyl or 3-(m-benzoyl)-phenyl at position 4. The crystallographic structure of VIM-2 in complex with an inhibitor showed the expected binding between the triazole-thione moiety and the dinuclear centre and also revealed a network of interactions involving Phe61, Tyr67, Trp87 and the conserved Asn233. Microbiological analysis suggested that the potentiation activity of the compounds was limited by poor outer membrane penetration or efflux. This was supported by the ability of one compound to restore the susceptibility of an NDM-1-producing E. coli clinical strain toward several β-lactams in the presence only of a sub-inhibitory concentration of colistin, a permeabilizing agent. Finally, some compounds were tested against the structurally similar di-zinc human glyoxalase II and found weaker inhibitors of the latter enzyme, thus showing a promising selectivity towards MBLs.

Full text of DOI:10.1016/j.ejmech.2020.112720

European Journal of Medicinal Chemistry 208 (2020) 112720
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
4-Amino-1,2,4-triazole-3-thione-derived Schiff bases as metallo-
lactamase inhibitors
b-
,
,
,
Laurent Gavara a **, Laurent Sevaille a 1, Filomena De Luca b 2, Paola Mercuri c,
,
Carine Bebrone c 3, Georges Feller d, Alice Legru a, Giulia Cerboni b, Silvia Tanfoni b,
Damien Baud a 4, Giuliano Cutolo a 5, Benoît Bestgen a 6, Giulia Chelini b,
,
,
,
Federica Verdirosa b, Filomena Sannio b, Cecilia Pozzi e, Manuela Benvenuti e,
Karolina Kwapien f , Marina Fischer , Katja Becker , Jean-Marie Frere ,
,
g,
7
h
h
c
ꢀ
e
f
g
c
ꢁ
Stefano Mangani , Nohad Gresh , Dorothee Berthomieu , Moreno Galleni ,
Jean-Denis Docquier b ***, Jean-François Hernandez a
,
, *
a
ꢁ
ꢁ
ꢁ
Institut des Biomolecules Max Mousseron, UMR5247 CNRS, Universite de Montpellier, ENSCM, Faculte de Pharmacie, 34093, Montpellier Cedex 5, France
Dipartimento di Biotecnologie Mediche, Universita di Siena, I-53100, Siena, Italy
Laboratoire des Macromolecules Biologiques, Centre D’Ingenierie des Proteines-InBioS, Universite de Liege, Institute of Chemistry B6a, Sart-Tilman, 4000,
b
c
ꢀ
ꢁ
ꢁ
ꢁ
ꢁ
ꢀ
ꢀ
Liege, Belgium
d
ꢁ
ꢁ
ꢁ
ꢀ
ꢁ
ꢀ
Laboratoire de Biochimie, Centre D’Ingenierie des Proteines-InBioS, Universite de Liege, Allee Du 6 Août B6, Sart-Tilman, 4000, Liege, Belgium
e
f
ꢀ
Dipartimento di Biotecnologie, Chimica e Farmacia, Universita di Siena, I-53100, Siena, Italy
Laboratoire de Chimie Theorique, UMR7616, Sorbonne Universite, CNRS, 75252, Paris, France
Institut Charles Gerhardt, UMR5253, CNRS, Universite de Montpellier, ENSCM, 34296, Montpellier Cedex 5, France
ꢁ
ꢁ
g
ꢁ
h Biochemistry and Molecular Biology, Interdisciplinary Research Center, Justus Liebig University, D-35392, Giessen, Germany
a r t i c l e i n f o
a b s t r a c t
Resistance to b-lactam antibiotics in Gram-negatives producing metallo-b-lactamases (MBLs) represents
Article history:
a major medical threat and there is an extremely urgent need to develop clinically useful inhibitors. We
previously reported the original binding mode of 5-substituted-4-amino/H-1,2,4-triazole-3-thione
compounds in the catalytic site of an MBL. Moreover, we showed that, although moderately potent,
they represented a promising basis for the development of broad-spectrum MBL inhibitors. Here, we
synthesized and characterized a large number of 4-amino-1,2,4-triazole-3-thione-derived Schiff bases.
Compared to the previous series, the presence of an aryl moiety at position 4 afforded an average 10-fold
Received 22 April 2020
Received in revised form
2 August 2020
Accepted 3 August 2020
Available online 20 August 2020
Abbreviations: AMP, Ampicillin; CAZ, Ceftazidime; CFX, Cefoxitin; COL, Colistin;
DMSO, dimethylsulfoxide; DMEM, Dulbecco’s Modified Eagle Medium; EDTA,
Ethylene diamine tetraacetic acid; EGM, Endothelial cell Growth Medium; FCS, Fetal
Calf Serum; HEPES, 4-(2-Hydroxyethyl)-1-piperazine-ethanesulfonic acid; hGloII,
human Glyoxalase II; HUVEC, Human Umbilical Vein Endothelial Cells; IPM, Imi-
penem; ITC, IsoThermal Calorimetry; LC-MS, liquid chromatography coupled to
mass spectrometry; LDH, Lactate DeHydrogenase; MBL, metallo-b-lactamase; MEM,
Meropenem; MHB, Mueller-Hinton broth; MIC, Minimum Inhibitory Concentration;
MOPS, 4-Morpholinopropane sulfonate; MTT, 3-(4,5-Dimethylthiazol-2-yl)-2,5-
diphenyltetrazolium bromide; TFA, trifluoroacetic acid.
* Corresponding author.
** Corresponding author.
*** Corresponding author.
E-mail addresses: jean-francois.hernandez@umontpellier.fr (J.-F. Hernandez),
laurent.gavara@umontpellier.fr (L. Gavara), jddocquier@unisi.it (J.-D. Docquier).
1
Institut Curie, Orsay, France.
GSK Vaccines Institute for Global Health, Siena, Italy.
2
3
ꢀ
ꢁ
^
ꢀ
Liege Universite, GIGA, CHU B34, 11 avenue de l’Hopital, 4000 Liege, Belgium.
Selvita, Poznan, Poland.
4
5
6
7
PrattLab, University of Southern California, Los Angeles, CA, USA.
Medicines for Malaria Venture, Geneva, Switzerland.
Sygnature Discovery, Nottingham, United Kingdom.
https://doi.org/10.1016/j.ejmech.2020.112720
0223-5234/© 2020 Elsevier Masson SAS. All rights reserved.

2
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
increase in potency. Among 90 synthetic compounds, more than half inhibited at least one of the six
tested MBLs (L1, VIM-4, VIM-2, NDM-1, IMP-1, CphA) with Ki values in the mM to sub-mM range. Several
Keywords:
Metallo-b-Lactamase
were broad-spectrum inhibitors, also inhibiting the most clinically relevant VIM-2 and NDM-1. Active
compounds generally contained halogenated, bicyclic aryl or phenolic moieties at position 5, and one
substituent among o-benzoic, 2,4-dihydroxyphenyl, p-benzyloxyphenyl or 3-(m-benzoyl)-phenyl at po-
sition 4. The crystallographic structure of VIM-2 in complex with an inhibitor showed the expected
binding between the triazole-thione moiety and the dinuclear centre and also revealed a network of
interactions involving Phe61, Tyr67, Trp87 and the conserved Asn233. Microbiological analysis suggested
that the potentiation activity of the compounds was limited by poor outer membrane penetration or
efflux. This was supported by the ability of one compound to restore the susceptibility of an NDM-1-
1,2,4-Triazole-3-thione
Schiff bases
Bacterial resistance
b-lactam antibiotic
producing E. coli clinical strain toward several b-lactams in the presence only of a sub-inhibitory con-
centration of colistin, a permeabilizing agent. Finally, some compounds were tested against the struc-
turally similar di-zinc human glyoxalase II and found weaker inhibitors of the latter enzyme, thus
showing a promising selectivity towards MBLs.
© 2020 Elsevier Masson SAS. All rights reserved.
1. Introduction
encountered difficulties, the subtle differences existing not only in
the active sites of MBLs from the B1 and B3 subclasses but also
The development and spread of bacteria resistant to antibiotics
is a very serious public health issue. The most relevant antibiotic
resistance mechanism in Gram-negative opportunistic pathogens is
the b-lactamase-mediated inactivation of b-lactam antibiotics, the
most widely used group of antibacterial agents, which includes
within a same sub-class render the discovery of a broad-spectrum
inhibitor a true challenge [8,16]. Also, most published inhibitors
possess a zinc-coordinating group, which represents a general
approach for the inhibition of metallo-enzymes. Therefore, another
difficulty is to avoid inhibition of some important human metallo-
enzymes.
penicillins, cephalosporins and carbapenems [1e4].
b
-Lactamases are grouped into four molecular classes: A, B, C,
Numerous families of MBL inhibitors have already been
described. The largest family of compounds possesses a thiol group
[e.g.17,18, which can be found in some marketed inhibitors of hu-
man metallo-peptidases (e.g. captopril, thiorphan). Whereas this
group could afford highly potent inhibitors because of its strong Zn-
binding properties, the consequence is a high risk for a lack of
selectivity. Another important family of MBL inhibitors contains
compounds with at least two carboxylate groups, such as succinate
analogues [19], which were not further developed, or as aspergil-
lomarasmine [20], a natural compound, whose inhibition mecha-
nism at least partially resembles that of ethylenediaminetetraacetic
acid (EDTA) [21]. Lack of selectivity could also be questioned for
these compounds as, for instance, aspergillomarasmine was found
to significantly inhibit ECE (Endothelin Converting Enzyme), an
essential metallo-peptidase involved in blood pressure control [22].
Another Zn-chelating MBL inhibitor, Zn148, was also recently re-
ported [23]. Finally, a more promising family is that of cyclic bor-
onates, which are large spectrum inhibitors of both MBLs and SBLs
(e.g. taniborbactam and QPX7728) [10e13,24,25]. Whereas
extended studies revealed a favourable selectivity, the capacity of
the boronate group to inhibit serine-, threonine- and metallo-
hydrolases increases the number of potential off-target activities.
In this context, we are studying compounds containing a 1,2,4-
triazole-3-thione moiety as an original zinc ligand. Otto Dideberg’s
group previously reported the 3D structure of L1, a di-zinc subclass
B3 MBL, with a published inhibitor, XIII ([26] Fig. 1) [27]. Surpris-
ingly, the active site only contained a fragment (IIIA, Fig. 1) pro-
duced via the hydrolysis of the hydrazone-like bond of XIII.
Whether the enzyme itself catalyzed the hydrolysis remains to be
determined. Furthermore, the inhibitor established an as yet un-
reported double coordination of the two Zn ions through two
atoms of its 1,2,4-triazole-3-thione moiety (namely N2 for Zn1 and
S3 for Zn2). Starting from this observation, we undertook a large
medicinal chemistry program to further assess the potential of
these compounds as MBL inhibitors. We recently reported a first
series of compounds, which were IIIA analogues (Fig. 1) varying by
their substituent at position 5 and substituted or not at position 4
by a NH2 group [28]. In this series, quite modest activities were
and D [5], showing an important heterogeneity of substrate pro-
files. Classes A, C and D are serine enzymes in which a serine res-
idue plays a central role in the catalytic activity. Some of them (e.g.
OXA-48- and KPC-type enzymes) also hydrolyse carbapenems,
which are considered last resort antibiotics in the hospital setting.
By contrast, all class B enzymes, which require a metal co-factor
(one or two Zn ions(s)) for their activity, are all efficient carbape-
nemases, often characterized by an exceedingly broad substrate
profile. These metallo-b-lactamases (MBLs) are classified into three
subclasses: B1, B2, and B3, based on structural features [6]. The
dizinc subclass B1 includes the most prevalent enzymes (e.g. VIM-,
NDM-, and IMP-types) found in the clinical setting. However,
whereas mono-zinc subclass B2 MBLs bear a limited clinical rele-
vance, some subclass B3 di-zinc enzymes have also been recently
acquired by relevant opportunistic pathogens (e.g. AIM-1, SMB-1).
MBLs are produced by multi-to extremely drug-resistant clinical
isolates belonging to some of the most clinically relevant Gram-
negative bacterial species, causing a significant number of noso-
comial infections (e.g. Klebsiella pneumoniae, Enterobacter cloacae,
Escherichia coli, Pseudomonas aeruginosa, Acinetobacter baumannii).
Furthermore, MBL-producing isolates, mainly found in the noso-
comial setting, are also increasingly reported in isolates causing
community-acquired infections [7]. Due to the global dissemina-
tion of carbapenem-resistant MBL-producing isolates, these en-
zymes are now regarded as major therapeutic targets.
A
validated approach to overcome
resistance consists of combination therapy associating a
drug to a -lactamase inhibitor, which protects the former from
inactivation. Several -lactamase inhibitors are currently available
(clavulanate, tazobactam, sulbactam, avibactam, relebactam and
vaborbactam), but they only target some serine- -lactamases
b
-lactamase-mediated
b
-lactam
b
b
b
(SBLs) and are not active on MBLs [8]. In fact, the discovery and
development of MBL inhibitors is extremely challenging. To our
best knowledge, only one MBL inhibitor reached the stage of clin-
ical development (taniborbactam or VNRX-5133, see below)
[9e13], and there are only a few other lead compounds efficiently
restoring bacterial sensitivity to carbapenems [14,15]. Among the

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
3
R1
5
N
NH2
N
N
HN
N
N
HN
N
N
HN
N
R2
4
COOH
S
S
S
XIII
IIIA
Fig. 1. Structure of compound XIII (1), a 4-amino-1,2,4-triazole-3-thione-derived Schiff base [26], IIIA [27,28] and general structure of synthesized analogues.
obtained, but a few micromolar and broad-spectrum analogues (as
tested against VIM-2, VIM-4, NDM-1, IMP-1, and L1) were identi-
fied, supporting the potential of this scaffold. In addition, we re-
ported the X-ray structure of L1 in complex with the 4-H analogue
of IIIA, which bound as IIIA but, unexpectedly, in a reverse manner
[28]. Ab initio quantum chemical and polarizable molecular me-
chanics calculations of the complex formed by a triazole-thione
motif with the dizinc active sites of L1 and VIM-2 have also been
performed [29].
Meanwhile, it is noteworthy that, in addition to the original in
silico screening by Olsen et al. on L1 [26], several other virtual and
experimental library screenings, including fragment-based
screening, on IMP-1 [30], VIM-2 [31] or NDM-1 [32] also identi-
fied this scaffold as a relevant di-zinc binding ligand. Furthermore,
the resolution of the 3D structure of VIM-2 in complex with a small
triazole-thione fragment showed a similar binding as IIIA in L1
[27,31]. These additional observations supported a very strong
rationale for further investigation of these series. A special interest
of the triazole-thione scaffold is that, compared to usual Zn ligands
(e.g. thiol), it presents higher hindrance and lower flexibility
because of the double coordination and is not as strong a Zn2þ
ligand. This should offer a higher probability to spare human
metallo-enzymes. Indeed, dizinc enzymes are much less abundant
in humans than monozinc ones.
compounds exhibiting > 70% inhibition.
First, we found that compound XIII (1, Table 1) was indeed a
more potent inhibitor than its heterocyclic fragment IIIA against all
tested enzymes (for IIIA: Ki ¼ 108
inhibition at 100 M (VIM-2, NDM-1, IMP-1) [28]; no data for
CphA). In particular, XIII was a micromolar inhibitor of VIM-2 and
VIM-4. In the case of L1 (Ki ¼ 40 M, a value in the same range as the
IC50 (15 M) reported by Olsen et al. [26]), the observed difference
with IIIA (Ki ¼ 108 M) suggested that XIII bound first to the
mM (L1), 81
mM (VIM-4), or ꢀ 30%
m
m
m
m
enzyme, followed by Schiff base hydrolysis and accommodation of
the IIIA fragment in the binding site as determined by crystallog-
raphy [27]. It is not known if this scenario could be a general in-
hibition mechanism for all MBLs and XIII analogues.
Based on this result, we synthesized several series of 4-amino-
1,2,4-triazole-3-thione-derived Schiff bases (XIII analogues). Their
inhibitory potencies are listed in Tables 1e3. Table 1 shows data
obtained for XIII analogues where the substituent at position 4 was
varied. The effect of variation at position 5 is presented on Table 2,
and Table 3 shows the activities of compounds combining favour-
able 4- and 5-substituents.
- Variation at position 4 (Table 1): the o-toluyl substituent at
position 5 was kept unchanged and the o-benzoic moiety at
position 4 was substituted with diverse aryl or heteroaryl
groups.
We herein report on a new series of 1,2,4-triazole-3-thione
compounds based on Olsen’s XIII Schiff base analogue [26]
(Fig. 1). Compared to IIIA analogues [28], the addition of an aryl
moiety at the 4-position of the heterocycle led to more potent and
broad-spectrum inhibitors of the most clinically relevant MBLs,
including VIM-2, NDM-1 and IMP-1.
First, the absence of substituent on the aryl group (compound 2)
was deleterious to activity. All other compounds contained a mono-
or di-substituted phenyl ring, with the exception of compound 21
that contained an imidazole ring.
In the case of compounds possessing an aryl mono-substituted
by small chemical groups, replacement of the ortho-carboxylate
group with a hydroxyl (3) or a nitrile (5) led to inactive compounds
against all tested MBLs with the exception of VIM-2 and VIM-4,
respectively. Similarly to compound 3, compound 6 with an o-
hydroxynapht-1-yl group was also inactive. In contrast, a nitro
group at the ortho position (compound 4) was more favourable as 4
not only possessed similar activities as compound XIII against L1,
VIM-4 and NDM-1, but also higher potency against IMP-1 and
2. Results and discussion
2.1. Synthesis
The 4-amino-1,2,4-triazole-3-thione compounds, which bear
the substituent at position 5, were synthesized as previously
described [28]. Then, the substituent at position 4 was introduced
by reaction between the 4-amine group and diverse benzaldehydes
in the presence of acetic acid to yield Schiff bases 1e90. (see
Scheme 1)
particularly CphA (Ki, 5.0
mM). With the exception of the 4-
fluorinated analogue 7, which was inactive against all tested
MBLs, all other compounds with a small substituent at the para
position showed micromolar Ki values against at least one MBL.
Whereas 8 (p-CO2H) moderately inhibited IMP-1, VIM-2 and CphA,
9 (p-OH) and 10 (p-NO2) were potent inhibitors of VIM-4 with Ki
values in the low micromolar range.
More significant results were obtained with compounds bearing
disubstituted aryl moieties, which combined hydroxy, methoxy,
carboxylic and nitro groups. The presence of two hydroxyl groups
2.2. Evaluation of inhibitory potency toward purified MBLs
All compounds were tested against six representative MBLs,
namely, the subclass B1 enzymes VIM-2, VIM-4, NDM-1 and IMP-1,
the subclass B2 CphA and the subclass B3 L1. Some compounds
were also tested against VIM-1. Initially, testing was performed at
one concentration (100 or 200 mM) and Ki values were measured for

4
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
A
B
C
R1
S
O
O
a
NH2
+
N
N
H2N
NH2
N
H
N
H
R1
OH
HN
S
R1
R1
O
c
b
d
NH2
N
HN
O
N
N
NH2
R1
N
H
R1
OEt
HN
S
S
R1
R1
e
NH2
N
R2
+
N
HN
N
N
HN
N
R2
O
1 - 90
S
S
Scheme 1. Synthesis of 4-amino-1,2,4-triazole-3-thione derivatives. A. R1 s Aryl. B. R1 ¼ aryl. C. Formation of Schiff bases 1e90 (R2 ¼ aryl). Reagents and conditions: (a) neat, 160 ꢁC;
(b) hydrazine hydrate, neat, 120 ꢁC, sealed tube; (c) CS2, KOH, EtOH, reflux; (d) hydrazine hydrate, EtOH, 100 ꢁC, sealed tube; (e) AcOH, reflux.
- Variation at position 5 of XIII (Table 2): the o-benzoic moiety at
position 4 was kept unchanged and the o-toluyl substituent at
position 5 was replaced by diverse aryl or heteroaryl groups
directly attached to the triazole ring or through a one- or two-
atom spacer.
was favourable for inhibition of both VIM-4 and VIM-2. Among the
three corresponding isomers 11, 12 and 13 (respectively containing
a 2,4-, 2,5- and 2,3-dihydroxyphenyl moiety), 11 with a hydroxyl
group at the para position showed the best activity profile as it also
moderately inhibited L1 and CphA. A para-hydroxyl group was also
favourable in compounds 15 (methoxy group at the ortho position)
and 16 (carboxyl group at the meta-position). In particular, 15
showed a large spectrum of activity with micromolar activities
against four MBLs (i.e. L1, VIM-4, VIM-2, CphA) and measurable
activity against NDM-1. In contrast, 14, the reverse isomer of 15,
only showed significant activity against IMP-1. Finally, the two
reverse isomers 17 and 18 (hydroxyl and nitro groups at the 2- and
5-positions) mainly showed significant inhibitory potency against
CphA and were not or very poor inhibitors of VIM-type enzymes.
Attaching a second aryl group to the 4-aryl substituent yielded
interesting compounds, 19 and 20. Compound 19 with a para-
benzyloxy moiety was a broad-spectrum inhibitor with significant
activities against L1, VIM-4, IMP-1 and CphA. It is noteworthy that
compound 19 was the best inhibitor of IMP-1 in this series with a Ki
Many substituents at position 5 in IIIA analogues (“short” series)
were previously evaluated [28]. In these “short” series (i.e. only NH2
or H at the 4-position), halogenated or bulky bicyclic aryl substit-
uent at the 5-position were favourable for activity, whereas com-
pounds containing more hydrophilic groups (i.e. phenol or
heteroaryl) were less or not active. Overall, as observed for XIII (1)
compared to IIIA, all Schiff bases were better inhibitors than their
short counterparts. Furthermore, numerous compounds of this
series showed better activity profile than XIII with larger spectrum
and enhanced inhibitory potencies, including against NDM-1.
Simply removing the o-methyl group of XIII o-toluyl substituent
yielded compound 22, which not only exhibited similar micromolar
inhibitory potency against VIM enzymes but also improved inhi-
bition against L1, NDM-1 and CphA. Adding one halogen at the meta
(Cl in 23; Br in 26; CF3 in 28) or para (Cl in 24) position was well
tolerated by VIM-4 but not by VIM-2 and L1 (with the exception of
26). Interestingly, micromolar inhibitory potencies were measured
against NDM-1 for the mono-chlorinated 23 and 24 and for the CF3-
of 0.3 mM but was devoid of activity against VIM-2 and NDM-1.
Compound 20 with a meta-benzoyl group at the 3-position of the
4-aryl substituent showed the largest spectrum in this series with
micromolar to submicromolar potencies against L1, VIM-2, VIM-4
and IMP-1, while CphA was moderately inhibited.
bearing analogue 28 (Ki, 3.2 mM). The fluorinated compound 27 was
Finally, compound 21 bearing an imidazole substituent was
found inactive toward all tested MBLs.
poorly or not active on tested MBLs. Despite these mono-
halogenated compounds possessed a restricted spectrum, the
addition of a second chlorine substituent interestingly yielded the
broad-spectrum inhibitor 25. Indeed, this compound was a potent
inhibitor of L1, VIM-2, VIM-4 and CphA, while showing some ac-
tivity on NDM-1. The enhanced activity against NDM-1 that was
observed in the halogen series is in agreement with the results
obtained with the “short” counterparts of 23 and 28 [28].
Overall, it is noteworthy that, although XIII was a modest in-
hibitor of NDM-1, none of these analogues were found to better
inhibit this enzyme. In addition, among all substituents evaluated
at position 4 of the triazole ring, some were of greater interest,
including 2-carboxy-, 2,4-dihydroxy-, 4-benzyloxy-, and 3-(m-
benzoic)-phenyl groups. Finally, a few compounds (i.e. 4, 15, 17, 18)
inhibited the mono-zinc B2 enzyme CphA with Ki values in the
micromolar range, confirming that this family of compounds could
inhibit MBLs from all three sub-classes [28].
In contrast to the corresponding “short” analogues, which were
poorly or not active [28], the phenolic compounds 29e31 were

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
5
Table 1
Inhibitory activity of 4-amino-1,2,4-triazole-3-thione-derived Schiff bases 1e21 with R1 ¼ o-tolyl against various MBLs.
Cpd
Structure
R2
Ki (m mM)
M) or (% inhibition ata100 orb200
L1
VIM-4
1.6
VIM-2
3.0
NDM-1
110
IMP-1
58.0
CphA
120
1
40.0
2
3
NI
NI
20.0
20.0
(32%)a
(34%)a
(41%)a
70.0
NI
(38%)a
(62%)a
NI
4
30.0
2.4
NI
NI
11.0
5.0
5
6
NI
NI
6.5
NI
NI
NI
(60%)b
(50%)b
ND
NI
NI
NI
7
(30%)a
(59%)a
(43%)a
(56%)a
NI
NI
(60%)b
(30%)a
11.0
ND
8
(57%)a
1.3
11.0
NI
NI
31.0
9
NI
(37%)a
(43%)a
(35%)a
(59%)a
10
1.4
NI
NI
(continued on next page)

6
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
Table 1 (continued )
Cpd
Structure
R2
Ki (m mM)
M) or (% inhibition ata100 orb200
L1
VIM-4
0.4
VIM-2
7.0
NDM-1
NI
IMP-1
CphA
62.0
11
31.0
(45% at 50 mM)
12
(48%)a
1.6
11.0
(37%)a
(54%)a
(54%)a
13
14
(55%)a
(55%)a
(65%)a
(55%)a
5.2
(39% at 50
mM)
(63%)a
(68%)a
(58%)a
(32%)a
9.0
32.0
15
6.0
2.4
10.0
120
NI
4.0
16
17
13.0
1.4
2.0
NI
120
NI
(45%)a
11.0
(39%)a
(61%)a
(44%)a
3.0
18
16.0
(32%)a
NI
NI
(47%)a
9.0
19
20
21
12.0
0.9
NI
2.1
0.3
NI
NI
NI
NI
NI
0.3
9.0
NI
22.0
1.10
40.0
NI
(36%)a

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
7
NI: no inhibition (<30% or < 50% inhibition at 100 or 200
m
M, respectively). ND: not determined. Kinetics were monitored at 30 ꢁC by following the absorbance variation
observed upon substrate hydrolysis. Ki’s were determined when inhibition >70% or for selected compounds showing broad-spectrum inhibition, regardless of their inhibition
potency. Assays were performed in triplicate (S.D. ꢀ 15%).
a
% inhibition at 100
% inhibition at 200
m
m
M.
M.
b
Table 2, 35 (naphth-2-yl in 5) and 44 (naphth-2-yl-methyl in 5):
respectively, (i) 67 (2,4-dihydroxyphenyl in 4), 68 (p-benzylox-
yphenyl in 4), 69 (3-(m-benzoic)-phenyl in 4), and (ii) 84 (2,4-
dihydroxyphenyl in 4), 85 (p-benzyloxyphenyl in 4).
potent inhibitors of VIM enzymes with submicromolar to micro-
molar Ki’s. In particular, 31 with a 2-hydroxy-5-methoxy-phenyl
moiety exhibited a quite large spectrum (i.e. L1, VIM-2, VIM-4,
CphA). However, while 30 and 31 moderately inhibited NDM-1,
none was active against IMP-1.
However, some interesting observations could be noted. The
decrease in activity was mainly observed for L1 and VIM-2. In the
case of L1, only two compounds exhibited Ki values below 10 mM:
Compared to their “short” analogues [28], which were
completely inactive against all tested MBLs, the pyridine analogues
32 and 33 showed significant, although still modest, activities.
In the “short” series, we reported that bi-aromatic substituents
were highly favourable to MBL inhibition [28]. The same was true in
the Schiff base series. Whereas compound 38 (o-biphenyl) was
inactive and compounds 34 (naphth-1-yl), 36 (1-hydroxy-naphth-
2-yl) and 37 (indol-2-yl) only potently inhibited one or two MBLs,
all others exhibited a broad spectrum of activity with micromolar to
submicrolar Ki’s: all MBLs for compound 39 (m-biphenyl); all but
CphA for 35 (naphth-2-yl) and 40 (p-biphenyl, not tested against
CphA); L1, VIM-4, VIM-2, NDM-1 for 41 (p-benzyloxyphenyl). In
addition, some compounds could be tested on VIM-1 and 40 and 41
63 (2,4-dihydroxyphenyl in 5, 3-(m-benzoic)-phenyl in 4) and 65
(naphth-1-yl in 5, 2,4-dihydroxyphenyl in 4). Micromolar and
submicromolar inhibitory potencies were measured against VIM-2
for only six compounds: five possessed a 2,4-dihydroxyphenyl
substituent in 4 (i.e. 55, 57, 62, 72, 86) and the sixth, a 3-(m-ben-
zoic)-phenyl moiety (63). The corresponding parent analogues 11
and 20 (Table 1) also inhibited VIM-2 with a similar potency. All
compounds containing a p-benzyloxyphenyl group in position 4
were inactive against this enzyme. This result confirmed that ob-
tained for compound 19 (Table 1), suggesting that the VIM-2
binding site cannot accommodate this bulky substituent.
significantly inhibited this enzyme with Ki values of 7.1 and 3.9 mM,
By contrast, VIM-4 was more tolerant, as already noticed in the
respectively.
parent series (Tables
1 and 2). Indeed, fifteen compounds
We also synthesized analogues where the aryl substituent was
separated from the triazole ring by one or two atoms. Whereas the
benzyl (42) and phenethyl (45) derivatives only significantly
inhibited one or two MBLs, all biaryl analogues (43, benzhydryl; 44,
naphth-2-ylmethyl; 46, benzhydrylmethyl; 47, naphth-2-
yloxymethyl) exhibited a broad spectrum of activity, all with Ki
composed of 14 different 5-subtituents and the three 4-
substituents (2,4-dihydroxyphenyl: 49, 51, 55, 57, 59, 61, 62, 64,
72, 76; p-benzyloxyphenyl, 54, 66, 87; 3-(m-benzoic)-phenyl: 63,
69) exhibited micromolar or submicromolar Ki values ranging
0.09e13
mM.
Although the parent XIII analogues containing
a
2,4-
values below 5 mM against NDM-1.
dihydroxyphenyl (11) or a p-benzyloxyphenyl at position 4 (19)
were inactive against NDM-1 (Table 1), eight compounds contain-
ing these two substituents potently inhibited this enzyme. This
underlines the importance of the substituent at position 5, rather
than at position 4, in providing better NDM-1 inhibitors. It is
noteworthy that, with the exception of compound 72 (o-biphenyl in
5), all other compounds contained 5-aryl substituents, which were
previously shown to be favourable (Table 2 and [28]), i.e. haloge-
nated aryl (50, 52, 53, 59) or biaryl ones (75, 76, 79). In contrast,
although 5-arylalkyl-containing compounds with an o-benzoic
group at position 4 (43, 44, 46, 47) were potent NDM-1 inhibitors
(Table 2), the related compounds 80e90 were inactive (Table 3).
Compared to compounds containing an o-benzoic group at the
4-position (Table 2), the results presented in Table 3 showed that, as
Therefore, by modifying the 5-substituent of XIII (1), we were
able to increase the inhibitory potency against NDM-1, up to 80-
fold for compound 40, the best NDM-1 inhibitor reported in this
paper (Ki, 1.3 mM). Although less important, an increase in potency
was also observed against all other MBLs. Compared to other en-
zymes, IMP-1 was more sensitive to changes as most compounds
were found inactive, but a few analogues were better IMP-1 in-
hibitors than XIII: the biphenyl analogues 39 and 40, and the
naphth-2-yl 35 and 44. Overall, several compounds showed a broad
spectrum of inhibition (Ki < 10 mM toward several enzymes): in
particular, two compounds, 35 and 39, showed Ki values in the
submicromolar or low micromolar range against 5 out of 6 tested
MBLs, which included two most clinically relevant MBLs NDM-1
and VIM-2. Compound 39 was also the best VIM-4 inhibitor re-
ported in this paper (Ki, 40 nM).
already observed for compound 19 (Ki, 0.3
zyloxyphenyl moiety at position 4 was highly favourable to IMP-1
inhibition, as Ki values ranging 0.1e11.4 M were measured for
mM; Table 1), a p-ben-
m
- Combinations of substituents at positions 4 and 5 selected from
the two preceding series (Table 3): selected substituents at po-
sition 5 (see Table 2) were combined with one, two, or all of the
following 4-substituent(s) considered as the most favourable
(see Table 1): 2,4-dihydroxyphenyl, p-benzyloxyphenyl and 3-
(m-benzoic)-phenyl. We also synthesized the analogue without
substituents on the two phenyl rings, 48, which was inactive
against all MBLs, except modestly so against IMP-1. Overall,
removal of the carboxylate group on the 4-subtituent phenyl
was generally detrimental to activity. For instance, in the case of
the most interesting broad-spectrum inhibitors possessing an o-
benzoic group at position 4 (i.e. 22, 25, 35, 39, 40, 43, 44, 46, 47,
Table 2), its substitution led to a decrease in inhibitory potency
against several if not all MBLs (respectively, 49, 53, [67e69], 74,
77, [82, 83], [84, 85], [88, 89], 90). The most drastic effect was
observed with the analogues of the two best inhibitors from
compounds 54, 66, 75, 77, 79 and 87. Three other molecules with
either a 3-(m-benzoic)-phenyl (63) or a 2,4-dihydroxyphenyl
moiety (65, 72) also significantly inhibited this enzyme.
Finally, among 17 tested compounds, seven exhibited Ki values
below 10 mM (i.e. 62, 64e66, 81, 86, 87) against the subclass B2
enzyme CphA. All were better inhibitors than their related ana-
logues with an o-benzoic group at the 4-position (30, 31, 34, 42, 45,
Table 2).
Overall, whereas replacing the o-benzoic group at the 4-position
was generally detrimental to inhibitory potencies and led to a
narrower spectrum profile, a few compounds without this sub-
stituent nonetheless exhibited potent activity against at least one
MBL. For instance, compounds 63 (Ki, 0.13
0.1 M) and 87 (Ki, 1.0 M) were respectively the best VIM-2, IMP-1
and CphA inhibitors reported in this paper. In addition, compounds
mM), 54 and 66 (Ki,
m
m

8
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
Table 2
Inhibitory activity of 4-amino-1,2,4-triazole-3-thione-derived Schiff bases 22e47 with R2 ¼ o-benzoic against various MBLs.
Cpd
Structure
R1
Ki (
m
M) or (% inhibition ata100 orb200
m
M)
VIM-2
L1
VIM-4
1.6
NDM-1
110
IMP-1
58.0
CphA
120.0
1
40.0
3.0
22
23
24
5.0
1.4
2.7
NI
38.0
7.5
NI
58.0
ND
(60%)a
(60%)a
2.0
7.0
58.0
NI
NI
7.6
ND
25
3.0
0.8
3.0
30.0
(53%)a
4.0
26
27
28
29
3.0
2.0
(35%)a
NI
NI
3.2
NI
NI
ND
(40%)a
(40%)a
(40%)a
(51%)a
NI
NI
ND
1.6
NI
(50%)b
(35%)a
3.0
2.2
NI
ND
30
38.0
0.2
0.6
79.0
NI
52.0

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
9
Table 2 (continued )
Cpd
Structure
R1
Ki (
L1
m
M) or (% inhibition ata100 orb200
m
M)
VIM-4
0.5
VIM-2
NDM-1
27.0
IMP-1
NI
CphA
13.0
31
9.0
0.7
32
(32%)a
(73%)a
(50%)b
(50%)b
(45%)a
NI
33
34
NI
(62%)a
(53%)a
NI
(60%)b
76.0
NI
NI
1.1
14.5
(35%)a
(56%)a
35
36
4.62
0.5
0.4
2.0
0.7
NI
2.3
10.6
NI
(60%)b
NI
ND
37
38
(45%)a
NI
NI
NI
9.4
54.7
ND
ND
NI
(49%)a
(32%)a
NI
39
40
0.6
0.9
0.04
4.0
0.2
2.5
4.2
1.3
17.7
32.4
4.0
ND
(continued on next page)

10
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
Table 2 (continued )
Cpd
Structure
R1
Ki (
m
M) or (% inhibition ata100 orb200
m
M)
VIM-2
L1
VIM-4
1.0
NDM-1
2.8
IMP-1
(62%)a
CphA
ND
41
0.4
0.9
42
43
(59%)a
0.1
(57%)a
9.0
3.2
(60%)b
4.1
(48%)a
NI
(51%)a
13.0
(50%)b
44
45
0.5
4.0
5.0
2.4
7.0
1.7
NI
34.4
15.0
(56%)a
(37%)a
(50%)a
46
2.4
5.6
25.4
2.7
NI
15.0
47
(60%)a
3.0
2.0
2.4
(35%)a
(42%)a
NI: no inhibition (<30% or < 50% inhibition at 100 or 200 m
M, respectively). ND: not determined. Kinetics were monitored at 30 ꢁC by following the absorbance variation
observed upon substrate hydrolysis. Ki’s were determined when inhibition >70% or for selected compounds showing broad-spectrum inhibition, regardless of their inhibition
potency. Assays were performed in triplicate (S.D. ꢀ 15%).
a
% inhibition at 100
% inhibition at 200
m
m
M.
M.
b
75 and 76 showed close Ki values (1.8
mM) to that of the best NDM-1
2.3. Isothermal calorimetry experiments (ITC)
inhibitor 40. Finally, this series afforded only two broad-spectrum
inhibitors: (i) 63 with potent activities against L1, VIM-4, VIM-2
and IMP-1, but not NDM-1, as the related 20 (Table 1); (ii) 72, which
more moderately inhibited VIM-2, VIM-4, IMP-1, but also NDM-1.
The binding of selected compounds to VIM-2 was studied by ITC.
Most inhibitors bound with an experimental stoichiometry n of
1.2e1.3, which clearly indicated a 1:1 association with the enzyme,
consistent with X-ray crystallography data (see below). Accord-
ingly, the compound concentration was lowered in the fitting
procedure in order to provide n z 1 and to take account of com-
pound solubility under ITC conditions. Typical ITC thermograms as
well as data analyses by a Wiseman plot are shown for compound
63 in Fig. 2 and compound 16 in Fig. 1S.
- The thione group is essential to MBL inhibition: some com-
pounds (a-c) where the sulphur atom was replaced by an oxy-
gen (1,2,4-triazole-3-one scaffold) were synthesized and tested
for inhibition (Scheme 1S and Table 1S). They shared the same 5-
substituent (phenyl ring) and differed by their 4-substituent. a
(o-benzoic),
b
(2,4-dihydroxyphenyl) and
c
(p-benzylox-
The thermodynamic parameters of binding are given in Table 4
and are also depicted in Fig. 3 for a more visual inspection (general
explanations are provided in Supplementary Information). As
shown, enthalpy drove the binding of most compounds (the
yphenyl) were all inactive or only poorly active against both L1,
VIM-4, VIM-2, NDM-1 and IMP-1 (Table 2S), confirming the
crucial role of the sulphur in binding.
enthalpic term is the major contributor to
D
Gꢁ and to affinity).
b

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
11
Table 3
Inhibitory activity of 4-amino-1,2,4-triazole-3-thione-derived Schiff bases 48e90 with combined R1 and R2 against various MBLs.
Cpd
Structure
R1
Ki (
L1
m mM)
M) or (% inhibition ata100 orb200
R2
VIM-4
NI
VIM-2
NI
NDM-1
NI
IMP-1
(60%)a
CphA
NI
48
NI
49
(46%)a
1.7
NDc
NDc
(54%)a
26.0
ND
50
51
(60%)a
(30%)a
(35%)a
NI
6.7
NI
3.0
(60%)b
(50%)b
(66%)a
(35%)a
52
53
(40%)a
(54%)a
NI
NI
4.1
3.1
NI
ND
NI
(63%)a
(61% at 50
mM)
(67%)a
54
55
18.0
0.6
4.0
NI
ND
0.1
NI
(40%)a
3.8
(50%)a
(55%)a
ND
56
57
(15%)a
NI
NI
NI
NI
NI
ND
ND
NI
13.0
0.62
(45%)a
58
59
NI
NI
NI
NI
NI
(60%)a
ND
ND
(25%)a
4.0
11.7
NI
60
NI
NI
NI
(50%)b
NI
ND
(continued on next page)

12
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
Table 3 (continued )
Cpd
Structure
R1
Ki (m mM)
M) or (% inhibition ata100 orb200
R2
L1
VIM-4
3.5
VIM-2
(53%)a
NDM-1
32.5
IMP-1
66.2
CphA
ND
61
(15%)a
62
(31%)a
0.5
2.0
NI
NI
6.0
63
64
2.7
0.09
3.0
0.13
NI
NI
NI
3.0
31.0
5.0
15.0
(50%)a
65
2.0
(32%)a
NI
NI
10.0
4.0
66
67
30.0
NI
2.4
NI
NI
NI
NI
0.1
NI
4.0
ND
(55%)a
68
NI
NI
NI
NI
NI
ND
69
70
(45%)a
NI
2.8
(30%)a
NI
(50%)b
NI
NI
NI
ND
ND
(55%)a
71
72
NI
NI
NI
NI
NI
(60%)b
12.0
ND
ND
9.0
4.5
6.4
73
NI
NI
NI
(45%)a
NI
ND

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
13
Table 3 (continued )
Cpd
Structure
R1
Ki (
L1
m mM)
M) or (% inhibition ata100 orb200
R2
VIM-4
(65%)a
VIM-2
(45%)a
NDM-1
(65%)a
IMP-1
NI
CphA
ND
74
NI
75
76
NI
NI
NI
NI
NI
1.8
1.8
9.0
NI
ND
ND
2.0
77
78
NI
NI
NI
NI
NI
NI
(64%)a
(42%)a
11.4
ND
ND
(62%)a
79
80
NI
NI
NI
3.2
NI
2.2
ND
NI
(46%)a
(47%)a
(30%)a
(37%)a
81
82
(31%)a
NI
NI
NI
NI
NI
NI
NI
5.5
ND
NI
(50%)a
83
84
NI
NI
NI
NI
NI
(60%)b
(50%)b
(64%)a
ND
ND
(30%)a
NI
85
86
NI
NI
NI
NI
NI
ND
6.0
(51%)a
(54%)a
9.0
168
(45%)a
87
32.0
2.0
NI
(30%)a
9.0
1.0
(continued on next page)

14
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
Table 3 (continued )
Cpd
Structure
R1
Ki (m mM)
M) or (% inhibition ata100 orb200
R2
L1
VIM-4
NI
VIM-2
ND
NDM-1
ND
IMP-1
(35%)a
CphA
NI
88
(35%)a
89
90
NI
NI
NI
NDc
NI
(30%)a
(33%)a
(30%)a
(40%)a
(60%)a
(30%)a
NI
NI: no inhibition (<30% or < 50% inhibition at 100 or 200 m
M, respectively). ND: not determined. Kinetics were monitored at 30 ꢁC by following the absorbance variation
observed upon substrate hydrolysis. Ki’s were determined when inhibition >70%. Assays were performed in triplicate (S.D. ꢀ 15%).
a
% inhibition at 100
% inhibition at 200
Not determined because of non-linear kinetics.
m
m
M.
M.
b
c
Compounds 16, 62, and in particular 57 and 63 (the latter being
among the best inhibitors of VIM-2 with Ki values of 0.62 and
0.13 mM, respectively) displayed the highest enthalpy values. Un-
fortunately, these favourable interactions were accompanied by a
noticeable entropy penalty (tight binding and loss of freedom).
Nevertheless, the favourable enthalpy contribution is an indication
of specific interactions between binding partners and reflects
ligand specificity and selectivity. Measured Kd values were close to
Ki values (Tables 1 and 3). Interestingly, these four compounds did
not contain an o-benzoic group at the 4-position in contrast to all
other tested inhibitors, suggesting that this substituent was
responsible of a specific behaviour.
In fact, all other investigated compounds (22, 25, 31, 35, 39)
displayed a favourable entropic contribution, which was even the
main contributor to affinity for the entropy-driven binding of
compounds 22, 25 and 35. However, favourable interactions per-
formed upon binding were relatively weak as reflected by the
corresponding enthalpic contribution, and as a result, the affinity
remained moderate. Overall, this suggested that favourable in-
teractions occurring in the complex were replacing pre-existing
bonds established with the solvent water and therefore the net
enthalpic effect was reduced due to desolvation [33,34]. In the case
of these five compounds, Kd values were three-to five-fold lower
than Ki values, with the exception of 35 for which these values were
equal (Table 2).
Finally, the compounds displayed a nearly perfect enthalpy-
entropy compensation phenomenon as illustrated in Fig. 2S,
which basically revealed that all compounds bound via the same
molecular mechanism. The regression line with r2 ¼ 0.98 indicated
that ITC experiments have retrieved high-quality data. The
slope ¼ 1.05 of this plot indicated that enthalpy (with favourable
interactions) and entropy (tight binding or desolvation) equally
contributed to the overall free energy of binding.
In the context of compound optimization, a strong preference is
given to enthalpy-driven binding, with favourable entropic
contribution (or at least a negligible contribution). In this respect,
compound 39, which not only displayed the highest affinity toward
VIM-2 in this experiment but was also one of the two best broad-
spectrum inhibitors in the present study, emerged as a promising
candidate.
Fig. 2. Isothermal titration calorimetry of VIM-2 by compound 63 at 25 ꢁC. Upper
panel: exothermic microcalorimetric trace of compound injections into VIM-2 solution
(18.7 mM). Lower panel: Wiseman plot of heat releases versus molar ratio of injectant/
protein in the cell and nonlinear fit of the binding isotherm for n equivalent binding
sites. The binding enthalpy corresponds to the amplitude of the transition curve, Ka is
derived from the slope of the transition and the stoichiometry n is determined at the
transition midpoint.

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
15
Table 4
Thermodynamic parameters of compounds binding to VIM-2 at 25 ꢁC, listed by decreasing affinity.
Cpd
n
Ka (Mꢂ1
)
Kd (nM)
D
Gꢁb (kcal/mol)
D
Нꢁ (kcal/mol)
TD
Sꢁ (kcal/mol)
b
b
39
63
57
31
22
35
25
16
62
1.01 0.01
1.00 0.01
1.03 0.02
1.01 0.01
1.01 0.03
1.00 0.01
1.02 0.01
1.02 0.01
1.01 0.04
1.5 0.2 107
7.2 0.7 106
3.1 0.6 106
2.6 0.3 106
1.9 0.6 106
1.4 0.1 106
1.2 0.1 106
7.7 0.7 105
3.4 0.6 105
68
ꢂ9.8
ꢂ9.3
ꢂ8.8
ꢂ8.7
ꢂ8.5
ꢂ8.4
ꢂ8.2
ꢂ8.0
ꢂ7.5
ꢂ7.2 0.1
ꢂ13.7 0.1
ꢂ10.4 0.2
ꢂ6.3 0.1
ꢂ3.6 0.2
ꢂ3.3 0.1
ꢂ3.9 0.1
ꢂ9.1 0.2
ꢂ9.6 0.5
2.6
139
323
389
519
713
863
1304
2980
ꢂ4.4
ꢂ1.6
2.4
4.9
5.1
4.3
ꢂ1.1
ꢂ2.1
Furthermore, and considering the growing epidemiological
relevance of NDM-type carbapenemases, compounds showing Ki
values lower than 2
also evaluated for their potential activity on a NDM-1-producing
multidrug-resistant clinical isolate, which also produces other
lactamases, including the extended-spectrum -lactamase CTX-M-
15, using a broth microdilution method. Tested at a fixed concen-
tration of 32 g/mL, these compounds showed some effect,
mM on NDM-1 (compounds 40, 44, 75, 76) were
b
-
b
m
although limited, with carbapenems (up to 4-fold reduction of the
MIC value for 76) (Table 6). Ceftazidime MIC was not affected by any
MBL inhibitors likely due to the production of class A extended-
spectrum
b-lactamase. In the presence of avibactam, a potent in-
hibitor of serine-
b
-lactamases, resistance to ceftazidime (and thus
Fig. 3. VIM-2 binding energetics of compounds, displayed by increasing affinity. Data
are from Table 4 (see the general signification of thermodynamic signatures in Sup-
porting Information).
to ceftazidime-avibactam) relies on the production of the NDM-1
enzyme. Compound 76 proved slightly more effective when the
medium was supplemented with a sub-inhibitory concentration of
colistin (that did not affect MIC values on its own), a polymyxin
antibiotic well known for its membrane permeabilizing properties
[36]. In these conditions, carbapenem MIC values were further
reduced by a 2-fold factor. The effect of colistin was especially
remarkable (64-fold reduction of the MIC value) with ceftazidime-
avibactam, as the addition of both 76 and colistin was able to
restore full susceptibility to the combination.
Altogether, these experiments (disk diffusion assays with wild-
type and efflux-impaired strains and broth microdilution assays,
including with colistin) indicated that the accumulation of these
inhibitors in the bacterial periplasm might rely on both efflux
mechanisms and limited diffusion through the outer membrane,
explaining their limited activity in whole cell assays, despite the
submicromolar inhibitory activity observed on purified enzymes.
However, the encouraging potentiation effect of compound 76
with both meropenem and ceftazidime-avibactam (in the presence
of colistin) supports the potential of these compounds as a starting
point of further medicinal chemistry efforts to overcome current
limitations.
2.4. In vitro antibacterial synergistic activity
The ability of selected compounds to potentiate the activity of b-
lactam antibiotics was evaluated using several methods. First, a
simple disk diffusion assay was performed on isogenic MBL-
producing laboratory Escherichia coli strains, obtained by trans-
forming strain XL-1 Blue or LZ2310 (a triple knock-out mutant of
K12 in which genes encoding components of the major efflux
pumps NorE, MdfA and AcrA were inactivated) with derivatives of
the pLB-II plasmid in which the blaVIM-2, blaVIM-4, blaIMP-1, and
blaNDM-1 genes were cloned [35]. The latter strain was used to limit
the potential efflux of the compounds, which would prevent suf-
ficient accumulation in the periplasm. Compounds were tested
according to their inhibition profile and availability. In this assay,
some compounds were found able to significantly potentiate the
activity of the antibiotic on LZ2310-derivative strains, mostly with
the VIM-2-producing strain (Table 5). These include 1 and 22,
which interestingly restored susceptibility to the antibiotic (being
above the EUCAST resistance breakpoint, 19 mm), as a diameter of
the growth inhibition zone of 20 mm was measured. However, a
more limited effect was observed when these compounds were
tested on the wild-type strain XL-1 Blue, indicating that some in-
hibitors might behave as substrate of at least one of these efflux
pumps. Nonetheless, compounds 13 and, to some extent, 61 pro-
vided some detectable potentiation of cefoxitin on VIM-2-
producing laboratory XL-1 Blue strain normally producing efflux
pumps. Other compounds, despite being rather potent inhibitors,
did not show any antibiotic potentiation, even when tested at up to
2.5. Selectivity toward human glyoxalase II
A representative panel of compounds (1, 11, 19, 22, 35, 44, 75)
was tested against human glyoxalase II, a bi-zinc metallo-hydrolase
belonging to the MBL superfamily. Table 7 shows that all com-
pounds could inhibit human glyoxalase II with IC50 values between
320 mg and the efflux impaired strains, indicating that they might
5.9 mM (19) and 30 mM (44). Unfortunately, in spite of several ap-
not be able to efficiently cross the outer membrane and accumulate
in the periplasm at biologically relevant concentrations. This was
especially true for compounds tested on the VIM-4- and NDM-1-
producing strains. Compounds 17 and, to some extent, 29 showed
potentiation when tested on the IMP-1-producing strain, with
compound 17 showing a remarkable dose-dependent effect.
proaches, exact Ki could not be determined for the human enzyme.
However, although Ki and IC50 values can only be indirectly
compared, most tested compounds were found more potent in-
hibitors of some MBLs. In particular, the broad-spectrum MBL in-
hibitors 35 and 44 showed clear selectivity in favour of VIM-2, VIM-
4, NDM-1 and L1.

16
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
e
Table 5
10
30
m
m
g ampicillin (AMP) disks were used to test the VIM-4-producing strain.
g ceftazidime (CAZ) disks were used to test NDM-1- and IMP-1-producing
f
In vitro synergistic activity of selected compounds measured by disk diffusion assays.
strains.
Cpd
Quantity (
m
g)a Diameter of growth inhibition zone (mm) observed
with strains producing the following MBLb:
g
Not determined.
VIM-2 VIM-2 (XL-1)c VIM-4 NDM-1 IMP-1
Table 6
Antibacterial activity of various
selected compounds on a NDM-1-producing E. coli clinical isolate determined by the
CFXd
AMPe
CAZf
1
14
-
8
e
6
e
e
6
e
e
e
9
e
e
9
b
-lactam antibiotics in the absence and presence of
g
e
e
e
9
e
e
15
e
broth microdilution method.
40
240
40
40
40
40
40
40
80
40
40
40
40
80
40
40
40
320
80
40
40
320
40
40
40
40
320
80
40
40
40
40
120
40
40
40
40
40
40
40
40
40
40
120
40
40
40
40
80
80
80
80
320
80
80
40
40
40
17
20
e
Cpd (32
m
g/mL)
MICa
IPM
(m )
g/mL) for E. coli (blaNþDM-1
e
4
8
10
11
12
13
e
e
e
e
e
e
15
e
e
e
e
e
e
e
e
8
e
e
e
10
e
e
e
e
9
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
12
e
e
e
e
e
e
e
e
10
e
e
e
e
e
6
e
e
MEM
CAZ
ZAV
14
e
e
9
6
6
6
e
e
e
7
e
e
e
e
e
e
9
e
e
13
19
9
None
40
44
16
8
8
8
8
64
32
64
64
16
8
>256
>256
>256
>256
>256
>256
>256
>256
>256
>256
>256
4
14
14
16
18
e
e
e
e
75
e
76
76 þ 0.12
a
g/mL COLb
4
14
15
16
17
e
m
14
14
e
e
IPM, imipenem; MEM, meropenem; CAZ, ceftazidime; ZAV, ceftazidime/avi-
g/mL).
Tested in the presence of sub-inhibitory concentration of colistin (COL).
6
e
bactam (avibactam was tested at a fixed concentration of 4
m
e
e
6
e
b
e
e
19
20
22
23
24
25
28
29
30
31
35
37
39
40
41
42
43
44
e
e
14
20
14
e
6
e
9
e
e
e
e
6
e
e
6
e
2.6. Cytotoxicity assays on mammalian cells
15
15
e
9
e
e
e
12
e
e
9
9
9
e
e
e
9
The potential cytotoxicity of compound XIII (1) was assessed
against HUVEC (human umbilical vein endothelial cells) using the
MTT assay. No toxicity was observed up to 100 mM concentration
(Fig. 3S). Furthermore, compound 76 did not induce cell lysis using
14
e
15
e
17
14
14
15
14
15
15
14
14
14
14
16
14
14
14
e
e
6
e
a membrane integrity assay (HeLa cells) at concentrations up to
e
e
e
e
e
e
e
e
e
e
e
e
6
e
e
e
6
e
e
e
6
15
15
e
250
mM. This result was also confirmed using a cell viability assay
(HeLa cells, 1500 cells/well), in which no cytotoxic effects could be
observed after up to 72 h of incubation in the presence of 250 mM
76.
16
15
e
15
15
e
9
2.7. Crystal structure of the VIM-2/cpd 31 complex
e
e
e
e
e
e
e
e
9
e
e
e
e
9
10
9
9
e
e
e
11
e
9
45
46
47
49
50
52
53
54
55
59
61
62
63
65
66
71
72
75
76
77
79
85
86
87
89
e
15
15
e
The three-dimensional structure of VIM-2 inhibited by com-
pound 31, one of the most potent VIM-2 inhibitor described herein,
was obtained at 1.95 Å resolution (Fig. 4, PDB code, 6YRP). The
binding of the inhibitor did not alter the tertiary structure of the
e
15
15
15
e
e
e
enzyme (RMSD, 0.176 Å: 210 Ca atoms). The structure showed that
e
the triazole-thione moiety similarly established a simultaneous
double coordination of the active site zincs, displacing the catalytic
hydroxide ion, as previously reported for short triazole-thione
compounds in the active site of L1 [27] (Fig. 4D) or VIM-2 [31],
the nitrogen at position 2 coordinating Zn1, while Zn2 was inter-
acting with the inhibitor sulphur atom. However, in contrast to
what was observed in the study by Nauton et al. [27] (i.e. IIIA, a
hydrolysis-derived fragment of XIII (1), was present in the L1 active
site), the entire Schiff base 31 was clearly present in the VIM-2
active site (as revealed by the electron density map, Fig. 4B),
showing that hydrolysis of the hydrazone-like bond did not occur
in this case. Compound 31 interacted with several active site resi-
dues (Fig. 4C). In particular, the methoxy and hydroxy oxygen
atoms of the substituent at position 5 were H-bonded to the NH of
Trp87 and NH2 of the conserved Asn233, respectively. Surprisingly,
the carboxylate moiety did not interact with Arg228 but was rather
H-bonded to the backbone NH of Asn233 and further interacted
with two water molecules. Another water molecule was found to
form an H-bond with the N1 atom of the triazole ring. The side-
chain of Arg228 was slightly displaced outwards the active site to
accommodate the inhibitor. Finally, the aromatic rings also inter-
acted with the side chains of Phe61 and Tyr67 with the establish-
14
e
e
15
e
e
14
14
e
e
6
e
e
e
e
6
e
e
e
e
e
e
e
e
e
e
6
e
14
e
15
15
15
e
14
16
e
15
e
e
14
e
e
e
e
9
e
e
e
15
a
Quantity of compound, dissolved in DMSO, deposited on the antibiotic-
containing disks. DMSO was used as a control and did not affect the diameter of
the growth inhibition zone. 220
mg EDTA were used as inhibition control and
restored full susceptibility to the antibiotic.
b
MBL-producing strains were obtained by transforming the pLB-II plasmid car-
rying the cloned MBL gene in strain LZ2310 (triple knock-out mutant of efflux
pumps), unless otherwise specified.
c
Strain obtained by transforming E. coli XL-1 Blue with the pLBII-VIM-2 plasmid,
and used to assess the potential activity of the compound in a wild type strain, i.e. in
ment of T-shaped
p stacking interactions and, for the benzoate
which the efflux pumps are normally produced.
d
30 mg cefoxitin (CFX) disks were used to test VIM-2 producing strains.
moiety, with the conserved His263 residue through parallel
p

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
17
Table 7
Inhibitory activities of compounds against human glyoxalase II.a
Cpd
Structure
IC50
(m
M)
Cpd
Structure
IC50 (mM)
1
25.8
35
13.2
11
19
19.0
44
75
30.0
23.4
5.9
22
25.5
a
GloII activity was measured by following the decrease in absorbance at
differed by less than 15%.
l
¼ 240 nm. All experiments were carried out in duplicate. Results are provided as mean values that
stacking interactions.
with an o-benzoic moiety. In particular, VIM-2 was more sensitive
Therefore, compound 31 interacted with several conserved
residues in the active site, including the catalytically important
Asn233, His263 and the hydrophobic residues Phe61, Tyr67 and
Trp87 (the latter also present in NDM-1), located in loop L1 and
defining one side of the active cavity. Overall, this structural in-
formation provides a structural basis to drive the synthesis of
further compounds.
to various substitution at this position than other enzymes,
whereas VIM-4 and NDM-1 were able to accommodate a larger
variety of analogues. IMP-1 showed a clear preference for the p-
benzyloxyphenyl substituent, that was not at all tolerated by VIM-
2, highlighting the substantial structural heterogeneity between
the active sites of these two MBLs. Thus, the 4-position appeared a
determining factor for selectivity. ITC experiments also highlighted
a
different binding behaviour between o-benzoic-containing
compounds and others. The former displayed a favourable entropic
contribution for binding while the later displayed the highest
enthalpy values.
X-ray crystallography confirmed the original binding mode of
triazole-thione with the dinuclear site of MBLs and provided a
structural basis for the potent inhibition of VIM-2 by 31, with the
establishment of a network of interactions with relevant residues,
such as the conserved Asn233. It also showed the surprising lack of
interaction of the inhibitor carboxylate with Arg228, which could
be exploited for the design of further inhibitors.
3. Conclusions
In conclusion, by varying the two substituents of compound XIII
(1) at positions 4 and/or 5, we were able to obtain highly potent
inhibitors of all tested MBLs. Compared to XIII, individual increases
in potency were approximately 400-, 40-, 23-, 85-, 580- and 120-
fold against L1 (43), VIM-4 (39), VIM-2 (63), NDM-1 (40), IMP-1
(54 and 66) and CphA (87), respectively. In addition, several com-
pounds showed broad-spectrum inhibition with Ki values in the
submicromolar or low micromolar range toward five out of six
representative MBLs, including two most clinically relevant en-
zymes such as VIM-2 and NDM-1. The most interesting inhibitors
contained an o-benzoic group at the 4-position and a biaryl moiety
at the 5-position (35, naphth-2-yl; 39, m-biphenyl; 44, naphth-2-
yl-methyl).
Twenty-seven different aryl groups attached at the 5-position of
the triazole ring were evaluated (see Table 2). The vast majority (i.e.
22) allowed potent inhibition of at least one MBL and half of these
afforded broad-spectrum inhibitors (Ki < 10.0
Twenty-one different aryl groups were evaluated at the 4-position
of the triazole ring (see Table 1). In this case, a lower tolerance was
Cytotoxicity experiments and the study of a potential off-target
(i.e. glyoxalase II) afforded quite favourable results. Microbiological
assays identified several compounds able to potentiate the activity
of a
b-lactam antibiotic when tested on several MBL-producing
laboratory strains, although there was a lack of correlation be-
tween their in vitro activity and inhibition potency. Some inhibitors
clearly showed a modest ability to accumulate in the periplasm at
significant concentrations, because of either active efflux or limited
diffusion through the outer membrane. However, we showed that
at least one potent NDM-1 inhibitor (76) was able to potentiate the
activity of carbapenems, and that this potentiation was further
improved in the presence of colistin, a polymyxin known for its
permeabilizing properties. This supports that additional efforts are
m
M for ꢃ 3 MBLs).
observed (only two compounds with Ki < 10.0
fact, higher potencies were generally measured for compounds
m
M for ꢃ 3 MBLs). In

18
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
group might be less straightforward and more likely to have an
impact on the inhibition potency.
Furthermore, one potential weakness of Schiff base analogues is
their susceptibility to hydrolysis, which would decrease compound
concentration while affording the moderately potent “short” ana-
logues. Although this hydrazone-like bond is quite stable in
aqueous solution, it might not be the case in biological media. In
support, L1 is suspected to cleave XIII into IIIA. Therefore, we are
now focusing on the synthesis of stable analogues either by
reduction of the double bond or by its substitution with a stable
spacer.
4. Experiment section
4.1. Chemistry section
4.1.1. General methods
NMR. 1H NMR and 13C NMR spectra were recorded at room
temperature on a 300, 400 or 500 MHz instrument in [D6]DMSO
solutions (unless otherwise indicated). Chemical shifts are reported
in parts per million (ppm) relative to residual solvent signals
(DMSO: 2.50 ppm for 1H and 39.52 for 13C). Splitting patterns in 1H
NMR spectra were designated as follows: s, singlet; d, doublet; t,
triplet; q, quartet; m, multiplet; br, broad. IR spectra were collected
on a PerkinElmer Spectrum One apparatus. Mp. Melting points
were determined on a capillary melting point apparatus. TLC. Thin-
layer chromatographies were performed on aluminium-backed
sheets of silica gel F254 (0.2 mm), which were visualized under
254 nm light, and by spraying with a 2% ninhydrin solution in
ethanol followed by heating, or by charring with an aqueous so-
lution of ammonium sulphate and sulfuric acid (200 g (NH4)2SO4
and 40 mL concentrated sulfuric acid in 1 L of water). Analytical
reverse phase HPLC. Samples were prepared in an acetonitrile/
water (50/50 v/v) mixture. Compounds insoluble in this mixture
were solubilized in DMSO. RP HPLC analyses were performed on a
Chromolith SpeedRod C18 column (0.46 ꢄ 5 cm) by means of a
linear gradient (0e100%) of 0.1% TFA/acetonitrile in 0.1% aqueous
TFA over 5 min, and at a flow rate of 3 mL/min. LC-MS. The LC/MS
system consisted of a Waters Alliance 2690 HPLC, coupled to a
Waters Micromass ZQ spectrometer (electrospray ionization mode,
ESIþ). All analyses were carried out using a RP C18 monolithic Onyx
Phenomenex 2.5 ꢄ 0.46 cm column. The flow rate was set to 3 mL/
min with eluent A (water/0.1% formic acid) and a gradient of 0%
/100% of eluent B (acetonitrile/0.1% formic acid) over 3 min was
then used. Positive ion electrospray mass spectra were acquired at a
Fig. 4. (A) Overall fold of VIM-2 in native form (PDB code, 1KO3; orange) and in
complex with compound 31 (JMV4690) (PDB code, 6YRP; green); (B) Close up of the
active site region of the VIM-2:31 complex (the omit map is shown in blue and con-
solvent flow rate of 100e500 mL/min. Nitrogen was used as both the
toured at approx. 2.5 s); (C) Mode of binding and interactions of 31 with VIM-2 active
site residues (green; native VIM-2 is shown in orange; see text for details); (D) Su-
perimposition of the L1:IIIA (PDB code, 2HB9) and VIM-2:31 complexes, showing the
similar positioning of the inhibitor in the dinuclear catalytic site. (For interpretation of
the references to color in this figure legend, the reader is referred to the Web version of
this article.)
nebulizing and drying gas. The data were obtained in a scan mode
in 0.1s intervals; 10 scans were added up to get the final spectrum.
HR-MS were registered on a JEOL JMS-SX-102A mass spectrometer.
Purification. (i) Column chromatography was performed using
Merck silica gel 60 of particle size 40e63 mm. (ii) Reverse phase
preparative HPLC purification of some final compounds was per-
worth pursuing to improve the penetration and accumulation in
the bacterial periplasm of 1,2,4-triazole-3-thione-based MBL in-
hibitors. We are now following various strategies known to have an
impact on penetration, besides working on expanding the scaffold’s
chemical space to generate useful SAR data allowing a better un-
derstanding of the requirements compatible with a better accu-
mulation of the inhibitor in the periplasm. For instance, potential
modifications could address the presence of a negative charge in
our most potent compounds potentially deleterious because of the
negatively charged lipopolysaccharides of the outer membrane by
introducing a positively charged group. Another possibility would
be the introduction of a siderophore group to exploit the bacterial
metal ion transport system, although the introduction of such
formed on a Waters Delta Pak C18 column (40 ꢄ 100 mm, 15
mm,
100 Å) by means of a linear gradient of eluent B in A at a 1%/min rate
(flow rate: 28 mL/min). Eluent A: water; eluent B: acetonitrile. TFA
was avoided because of susceptibility of the hydrazone-like bond to
acid-promoted hydrolysis during freeze-drying.
The purity of all compounds was determined to be ꢃ 95% by 1H
NMR and reverse phase HPLC.
The preparation of 4-amino-5-substituted-triazole-3-thiones
was performed as already reported [28] and is briefly described
in Supplementary Information.
4.1.2. General procedure for the preparation of Schiff bases
To a solution of 4-amino-1,2,4-triazole-3-thione (0.5 mmol, 1

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
19
eq.) in acetic acid (4 mL) was added a variously substituted ary-
performed in triplicate. Errors were less than 15%.
laldehyde (0.75 mmol, 1.5 eq.). The mixture was heated to reflux for
2 h, and then allowed to cool to room temperature. The precipitate
was filtered-off and recrystallized in ethanol to provide the corre-
sponding Schiff base.
4.2.1.2. CphA and L1. Compounds were prepared as 10 mM solu-
tions in DMSO before dilution with a 15 mM Cacodylate pH 6.5
buffer for CphA and 25 mM HEPES pH 7.5 for L1. L1 and CphA,
purified as previously described [27,41,42], were respectively tested
4.2. Biology section
with and without 50
were used at fixed concentrations between 0.2 and 0.8 nM. The
enzyme and the inhibitor (100 M) were pre-incubated for 30 min
in a volume of 495 l of 10 mM imi-
L at 21 ꢁC. Then, for CphA 5
penem substrate were added and its hydrolysis was monitored by
following the absorbance variation at 300 nm using a Specord 50
PLUS spectrophotometer (Analytik Jena, Germany). For L1, the re-
mM final ZnCl2 concentration. The enzymes
4.2.1. Metallo-b-Lactamase inhibition assays
m
4.2.1.1. VIM-type enzymes, NDM-1 and IMP-1. The inhibition po-
tency of the compounds has been assessed with a reporter sub-
strate method and specifically by measuring the rate of hydrolysis
m
m
of the reporter substrate (150
1; 120 M cefotaxime or 100
m
m
M imipenem for VIM-2, VIM-4, IMP-
m
M nitrocefin for NDM-1) at 30 ꢁC in
porter substrate was nitrocefin (100 mM final concentration) and its
50 mM HEPES buffer (pH 7.5) in the absence and presence of several
concentrations of the inhibitor (final concentration, 0.5e1000 M).
hydrolysis was followed at 482 nm. Experiments were performed
at 30 ꢁC. The activity was tested by measuring the initial rates in
three samples without inhibitor, which allowed determining the
m
The reaction rate was measured as the variation of absorbance
observed upon substrate hydrolysis in a UVeVis spectrophotom-
eter or microplate reader at a wavelength of 300 nm (imipenem,
meropenem), 260 nm (cefotaxime), or 482 nm (nitrocefin) and in
the presence of a purified MBL enzyme (such as VIM-2, VIM-4,
NDM-1, and IMP-1, at a final concentration ranging from 1 to
70 nM).
percentage of residual
hibitors. When the residual activity in the presence of 100
b
-lactamase activity in the presence of in-
M in-
m
hibitor was <30%, the initial rate conditions were used to study the
inhibition in the presence of increasing concentrations of com-
pounds (from 1 to 100 mM) and to determine the competitive in-
hibition constant, Ki. The Hanes linearization of the Henri-Michaelis
equation. was used and Ki was calculated on the basis of the
These enzymes were overproduced in derivatives of E. coli
BL21(DE3) carrying a pET-9a (IMP-1, VIM-2, VIM-4) or pET-15b
following equation:
were performed in triplicate. Errors were less than 15%.
n
¼ VmaxS/[S þ Km(1 þ I/Ki)] [18]. The assays
(NDM-1) vector in which the metallo-
reading frame was cloned. Bacterial strains were grown in the auto-
inducing medium ZYP-5052 supplemented with 50 g/mL kana-
b-lactamase-encoding open
m
4.2.2. Microbiological assays
mycin for 24 h at 37 ꢁC [37]. The IMP-1, VIM-2 and VIM-4 were
purified by ion-exchange and gel filtration chromatography using
an AKTA Purifier platform (GE Healthcare, Uppsala, Sweden) as
previously described [38]. NDM-1 was purified using affinity
chromatography. After growth for 26 h at 25 ꢁC, bacterial cells were
harvested by centrifugation (10,000ꢄg, 30 min, 4 ꢁC), resuspended
in 10 mM HEPES buffer (pH, 7.5) containing 0.15 M NaCl (buffer A)
in one fifth of the original culture volume and lyzed by sonication.
Cellular debris was removed by centrifugation (12,000ꢄg, 40 min,
4 ꢁC). The clarified sample was loaded on a HisTrap HP column (bed
volume, 5 mL; GE Healthcare), followed by a wash step (25 mL)
with buffer A containing 20 mM imidazole. Proteins were eluted
using buffer A containing 350 mM imidazole. The resulting protein
sample was desalted (HiPrep 26/10 Desalting column, GE Health-
care) using 20 mM triethanolamine (pH, 7.2) supplemented with
The potential synergistic activity of selected compounds was
assessed by both agar disk-diffusion and broth microdilution
methods, according to the CLSI (Clinical Laboratory Standard
Institute) recommendations, using Mueller-Hinton broth [43].
In the former, antibiotic-containing disks (30
ampicillin or 30 g ceftazidime) were supplemented with variable
amounts, according to both solubility and availability, of the in-
mg cefoxitin, 10 mg
m
hibitor (dissolved in DMSO, a maximum volume of 15
mL was
deposited on a disk). After incubation (18 h, 35 1 ꢁC), the diameter
of the growth inhibition zone was measured and compared to that
obtained in the absence of inhibitor. DMSO and EDTA (220 mg) were
used as negative and positive inhibition controls, respectively. Ex-
periments were performed in triplicate.
These tests were performed on MBL-producing isogenic labo-
ratory strains obtained after transforming Escherichia coli XL-1Blue
(Invitrogen, Carlsbad, USA) and LZ2310 strains with a derivative of
50 mM ZnSO4 and concentrated using an Amicon Ultra-15 (MW cut-
off, 10 kDa; Millipore, Burlington, Mass.). The concentrated sample
was loaded on a Superdex 75 prep grade column (XK16/100 col-
umn; bed volume, 140 mL; GE Healthcare) and the proteins eluted
the high copy number plasmid pLB-II carrying the cloned blaVIM-2,
blaVIM-4, blaNDM-1 and blaIMP-1 genes. Strain LZ2310 is a triple knock-
out mutant derived from K12 in which the genes encoding for
components of the three major efflux pumps systems were inac-
with buffer A containing 50 mM ZnSO4. The active fractions were
pooled and stored at ꢂ20 ꢁC. The purity of the protein preparations
was estimated >95% by SDS-PAGE analysis. The authenticity of the
enzyme preparations was confirmed by electron spray ionization
mass spectrometry, as previously described [39].
tivated (genotype, DnorE DmdfA N43 acrA1) [44] and was used to
generate a hyperpermeable background for MBL production and
investigate the potential contribution of such mechanisms to in-
hibitor efflux.
The inhibition constants (Ki) were determined on the basis of a
model of competitive inhibition by analysing the dependence of the
ratio v0/vi (v0, hydrolysis velocity in the absence of inhibitor; vi,
hydrolysis velocity in the presence of inhibitor) as a function of [I]
as already described [40]. The slope of the plot of v0/vi vs [I], which
corresponds to KmS /(KmS þ [S])Ki (where KmS is the Km value of the
reporter substrate and [S] its concentration in the reaction mixture)
and allowed the calculation of the Ki value. Alternatively, a Dixon
plot analysis was carried out by measuring the initial hydrolysis
rates in the presence of variable concentrations of inhibitor and
substrate. This allowed Ki values to be determined and supported
the hypothesis that the various compounds behaved as competitive
inhibitors of the various tested enzymes. The assays were
The minimum inhibitory concentrations (MICs) of imipenem
(IPM), meropenem (MEM), ceftazidime (CAZ) and the combination
CAZ-avibactam were determined in triplicate using Mueller-Hinton
broth and a bacterial inoculum of 5 ꢄ 104 CFU/well, as recom-
mended by the CLSI [45], in both the absence and presence of a
fixed concentration (32 mg/mL) of the inhibitor. A multidrug resis-
tant NDM-1-producing E. coli clinical isolate present in our collec-
tion (SI-001M) was used. This experiment was also performed in
the presence of a sub-inhibitory concentration of colistin (0.12
mL).
mg/
4.2.3. Inhibition of human glyoxalase II (hGloII)
The recombinant enzyme was produced in E. coli and purified as

20
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
described [46]. For measuring GloII activity, the decrease in
recommended by the supplier, using the above mentioned condi-
tions, except for cell density (1500 cells/well).
absorbance
resulting
from
S-D-lactoylglutathione
(e240nm ¼ 3.1 mMꢂ1 cmꢂ1) hydrolysis was measured at 25 ꢁC in
100 mM 4-morpholinopropane sulfonate (MOPS) buffer, pH 7.2 in a
total volume of 0.5 mL and with a Thermo Evolution spectropho-
4.2.5. Isothermal titration calorimetry (ITC) analysis of compound
binding to VIM-2
tometer. S-
D
-lactoylglutathione was obtained from Sigma. In-
ITC titrations were performed on a MicroCal ITC200 (GE-Mal-
hibitors were dissolved in DMSO and added to the assay mixture
containing buffer and substrate before the reaction was started
with enzyme (5e25 nM hGloII). For each assay, background ab-
sorbances induced by the inhibitors were taken into account. All
experiments were carried out in duplicate. Results are provided as
mean values that differed by less than 15%.
vern) equipped with a 200
mated 40 l glass syringe rotating at 1000 rpm. VIM-2 in 10 mM
HEPES-NaOH, 0.15 M NaCl, 50 M ZnSO4, pH 7.5 was diluted to the
desired concentration with the same buffer and was brought to
DMSO concentration identical to that of the injected compound.
The tested compounds were solubilized in DMSO at 20 mM con-
ml Hastelloy sample cell and an auto-
m
m
centration and were diluted to 200e400
buffer, resulting in a final DMSO concentration of 1e2%.
In a standard experiment, VIM-2 (19 M) was titrated by one
initiating injection (0.5 l) followed by 19 injections (2 l) of
compounds (200e400 M) at an interval of 150 s. Dilution heat of
mM with the enzyme
4.2.4. Cell toxicity
4.2.4.1. MTT assay. Cell culture: HUVEC were maintained in culture
in 10 cm plates with EGM-2 (endothelial cell growth media) con-
taining 10% fetal calf serum (FCS). After reaching confluence, these
cells were detached using trypsin and EDTA. The obtained cell
suspension was centrifuged and cells were counted and placed in
m
m
m
m
compound injections into buffer, at the corresponding DMSO con-
centration, were subtracted from raw data.
96-well culture plates at a density of 3000 cells/100
were allowed to adhere for 24 h.
m
L/well. Cells
The data so obtained were fitted via the non-linear least squares
minimization method to determine binding stoichiometry (n), as-
MTT assay: cell viability was assessed using the MTT colori-
metric assay. The yellow substrate MTT (3-(4,5-dimethylthiazol-2-
yl)-2,5-diphenyltetrazolium bromide) is reduced only by meta-
bolically active viable cells through the action of mitochondrial
dehydrogenases resulting in the formation of blue formazan crys-
tals, which do not cross the cell membrane and therefore accu-
mulate within viable cells. Solubilization of the crystals with DMSO
afforded a purple solution, which intensity is directly proportional
to the number of viable cells.
sociation constant (Ka), and change in enthalpy of binding (D )
Hꢁ
b
using ORIGIN 7 software v.7 (OriginLab). The Gibbs free energy of
binding,
entropic term, T
equation using the experimental
D
Gꢁb, was calculated from Ka
(D
Gꢁ ¼ -RT ln Ka) and the
b
D
Sꢁb, was derived from the Gibbs-Helmholtz
D
Hꢁb value (
D
Gꢁ
¼
D D
Hꢁb - T Sꢁb).
b
4.3. X-ray crystallography section
The VIM-2 -lactamase was purified and crystallized using the
b
To evaluate its cytotoxicity, HUVEC in 96-wells plates were
treated with compound XIII (1) (from a 10 mM stock solution in
sitting drop method essentially as previously described [15,40].
Crystals of the native enzymes were soaked for up to 60 min with a
5 mM inhibitor solution in the crystallization buffer, prior to flash-
freezing in the presence of ethylene glycol as the cryoprotectant. X-
ray diffraction data were collected at the European Radiation Syn-
chrotron Facility (ESRF, Grenoble, France). Data were processed as
described elsewhere [47,48] and the structures obtained by mo-
lecular replacement using the native structure of the corresponding
enzyme (PDB code 1KO3), excluding solvent and other ligands.
Refinement was performed with REFMAC5 [49] from the CCP4 suite
[50], using an iterative manual rebuilding and modelling of missing
atoms in the electron density using COOT [51]. Water molecules
were added using the standard procedures implemented in the
ARP/wARP suite [52]. Data collection and model refinement sta-
tistics are shown in Table 1S. The coordinates and structure factors
were deposited in the Protein Data Bank under code 6YRP.
DMSO) at increasing concentrations, 1, 10, 50 and 100
mM
(control ¼ EGM-2 without serum in the presence of 1% DMSO) after
removing the adhesion medium. Treated HUVEC were incubated at
37 ꢁC under 5% CO2 for 48 h. The MTT test was then performed.
100
ml of MTT solution (previously prepared in DMEM without
phenol red) were added to each well. After 4 h, cells were observed
under a microscope and the MTT solution was removed. After
adding 50 ml of DMSO the absorbance of the resulting solution was
measured using a microplate reader (TECAN) at a wavelength of
540 nm.
Average absorbance values (þ/ꢂ SD) were calculated from two
experiments done in triplicate (Fig. 3S).
4.2.4.2. Membrane integrity and cell viability assays. The potential
cytotoxic activity of compound 76 was evaluated using the
commercially available membrane integrity assay (CytoTox 96®
non-radioactive cytotoxicity assay, Promega, Madison, WI, USA).
The compound was tested for its ability to induce the lysis of HeLa
cells by measuring the release of lactate dehydrogenase (LDH) after
incubating the HeLa cell cultures (20,000 cells/well) for 24 h (37 ꢁC,
5% CO2) in DMEM (Dulbecco’s Modified Eagle Medium) supple-
mented with 10% fetal bovine serum, 4.5 mg/mL glucose and 2 mM
Declaration of competing interest
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
Acknowledgements
L-glutamine in the absence and presence of varying concentrations
of the compound (up to 250 M). Further controls included samples
m
Part of this work was supported by Agence Nationale de la
Recherche (ANR-14-CE16-0028-01, including fellowships to L.S. and
K.K.), the Deutsche Forschungsgemeinschaft (BE1540/15-2 within
SPP 1710 to K.B.). We thank Mr Pierre Sanchez for mass spec-
trometry analyses, Lucia Morbidelli and Vanessa Ritrovato for
cytotoxicity studies. Thanks are also due to Prof. Lynn Zechiedrich
(Baylor College of Medicine, Houston, USA) for providing the E. coli
strain LZ2310. CP, MB and SM thank the support of MIUR-Italy
“Progetto Dipartimenti di Eccelenza 2018-2022” allocated to
Department of Biotechnology, Chemistry and Pharmacy of the
containing the medium only or in which cell lysis was induced by
the addition of 9% Triton X-100 (maximum LDH release control).
The percentage of cytotoxicity was computed as 100 x (sample LDH
release)/(maximum LDH release). The variation of the percentage
of cytotoxicity allowed to compute an IC50 value (compound con-
centration inducing 50% cytoxicity). In addition, a commercial cell
viability assay (RealTime-Glo™ MT Cell Viability Assay, Promega)
was used to evaluate potential cytotoxicity effects upon longer in-
cubation times (up to 72 h). The assay was performed as

L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
21
Mercaptophosphonate compounds as broad-spectrum inhibitors of the met-
allo- -lactamases, J. Med. Chem. 53 (2010) 4862e4876.
[19] J.H. Toney, G.G. Hammond, P.M. Fitzgerald, N. Sharma, J.M. Balkovec,
G.P. Rouen, S.H. Olson, M.L. Hammond, M.L. Greenlee, Y.D. Gao, Succinic acids
University of Siena. CP, MB, SM, and JDD thank the European Syn-
chrotron Radiation Facility (Grenoble, France) for beamline access
and the staff of beamline ID23-1 for their kind assistance.
b
as potent inhibitors of plasmid-borne IMP-1 metallo-b-lactamase, J. Biol.
Chem. 276 (2001) 31913e31918.
[20] A.M. King, S.A. Reid-Yu, W. Wang, D.T. King, G. De Pascale, N.C. Strynadka,
T.R. Walsh, B.K. Coombes, G.D. Wright, Aspergillomarasmine A overcomes
Appendix A. Supplementary data
metallo-b-lactamase antibiotic resistance, Nature 510 (2014) 503e506.
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.ejmech.2020.112720.
[21] A. Bergstrom, A. Katko, Z. Adkins, J. Hill, Z. Cheng, M. Burnett, H. Yang,
M. Aitha, M.R. Mehaffey, J.S. Brodbelt, K.H. Tehrani, N.I. Martin, R.A. Bonomo,
R.C. Page, D.L. Tierney, W. Fast, G.D. Wright, M.W. Crowder, Probing the
interaction of aspergillomarasmine A with metallo-b-lactamase NDM-1, VIM-
2, and IMP-7, ACS Infect. Dis. 4 (2018) 135e145.
References
[22] A. Matsuura, H. Okumura, R. Asakura, N. Ashizawa, M. Takahashi, F. Kobayashi,
N. Ashikawa, K. Arai, Pharmacological profiles of aspergillomarasmines as
endothelin converting enzyme inhibitors, Jpn. J. Pharmacol. 63 (1993)
187e193.
[1] P. Nordmann, T. Naas, L. Poirel, Global spread of carbapenemase-producing
Enterobacteriaceae, Emerg. Infect. Dis. 17 (2011) 1791e1798.
[2] T.R. Walsh, M.A. Toleman, The emergence of pan-resistant Gram-negative
pathogens merits a rapid global political response, J. Antimicrob. Chemother.
67 (2012) 1e3.
€
[23] O. Samuelsen, O.A.H. Astrand, C. Frohlich, A. Heikal, S. Skagseth, T.J.O. Carlsen,
H.S. Leiros, A. Bayer, C. Schnaars, G. Kildahl-Andersen, S. Lauksund, S. Finke,
S. Huber, T. Gjoen, A.M.S. Andresen, O.A. Okstad, P. Rongved, ZN148 e a
modular synthetic metallo-b-lactamase inhibitor reverses carbapenem-
resistance in Gram-negative pathogens in vivo, Antimicrob. Agents Chemo-
[3] W.C. Reygaert, An overview of the antimicrobial resistance mechanisms of
bacteria, AIMS Microbiol. 4 (2018) 482e501.
€
[4] A. Cassini, L.D. Hogberg, D. Plachouras, A. Quattrocchi, A. Hoxha,
ther. (2020), https://doi.org/10.1128/AAC.02415-19.
[24] J. Brem, R. Cain, S. Cahill, M.A. McDonough, I.J. Clifton, J.C. Jimenez-Castellanos,
G.S. Simonsen, M. Colomb-Cotinat, M.E. Kretzschmar, B. Devleesschauwer,
M. Cecchini, D.A. Ouakrim, T.C. Oliveira, M.J. Struelens, C. Suetens, D.L. Monnet,
Burden of AMR Collaborative Group. Attributable deaths and disability-
adjusted life-years caused by infections with antibiotic-resistant bacteria in
the EU and the European Economic Area in 2015: a population-level model-
ling analysis, Lancet Infect. Dis. 19 (2019) 56e66.
ꢁ
M.B. Avison, J. Spencer, C.W. Fishwick, C.J. Schofield, Structural basis of met-
allo-b-lactamase, serine-b-lactamase and penicillin-binding protein inhibition
by cyclic boronates, Nat. Commun. 7 (2016) 12406.
[25] S.J. Hecker, K.R. Reddy, O. Lomovskaya, D.C. Griffith, D. Rubio-Aparicio,
K. Nelson, R. Tsivkovski, D. Sun, M. Sabet, Z. Tarazi, J. Parkinson, M. Totrov,
S.H. Boyer, T.W. Glinka, O.A. Pemberton, Y. Chen, M.N. Dudley, Discovery of
cyclic boronic acid QPX7728, an ultra-broad-spectrum inhibitor of serine and
metallo-b-lactamases, J. Med. Chem. (2020), https://doi.org/10.1201/
acs.jmedchem.9b01976.
[26] L. Olsen, S. Jost, H.W. Adolph, I. Pettersson, L. Hemmingsen, F.S. Jørgensen,
New leads of metallo-b-lactamase inhibitors from structure-based pharma-
cophore design, Bioorg. Med. Chem. 14 (2006) 2627e2635.
[27] L. Nauton, R. Kahn, G. Garau, J.-F. Hernandez, O. Dideberg, Structural insights
into the design of inhibitors of the L1 metallo-b-lactamase from Steno-
trophomonas maltophilia, J. Mol. Biol. 375 (2008) 257e269.
[28] L. Sevaille, L. Gavara, C. Bebrone, F. De Luca, L. Nauton, M. Achard, P. Mercuri,
S. Tanfoni, L. Borgianni, C. Guyon, P. Lonjon, G. Turan-Zitouni, J. Dzieciolowski,
[5] K. Bush, Past and present perspectives on
Chemother. 62 (2018) e01076-18.
b-lactamases, Antimicrob. Agents
[6] T. Palzkill, Metallo-
b-lactamase structure and function, Ann. N. Y. Acad. Sci.
1277 (2013) 91e104.
[7] V.R. Gajamer, A. Bhattacharjee, D. Paul, C. Deshamukhya, A.K. Singh,
N. Pradhan, H.K. Tiwari, Escherichia coli encoding blaNDM-5 associated with
community-acquired urinary tract infections with unusual MIC creep-like
phenomenon against imipenem, J. Glob. Antimicrob. Resist. 14 (2018)
228e232.
[8] J.-D. Docquier, S. Managani, An update on b-lactamase inhibitor discovery and
development, Drug Resist. Updates 36 (2018) 13e29.
[9] C. J. Burns, D. Daigle, B. Liu, D. McGarry, D. C. Pevear, R. E. Trout,
inhibitors. WO Patent WO 2014/089365 A1.
b-Lactamase
ꢁ
ꢀ
K. Becker, L. Benard, C. Condon, L. Maillard, J. Martinez, J.-M. Frere,
O. Dideberg, M. Galleni, J.-D. Docquier, J.-F. Hernandez, 1,2,4-Triazole-3-
[10] A. Krajnc, P.A. Lang, T.D. Panduwawala, J. Brem, C.J. Schofield, Will morphing
boron-based inhibitors beat the
b
-lactamases? Curr. Opin. Chem. Biol. 50
thione compounds as inhibitors of dizinc metallo-b-lactamase, Chem-
MedChem 12 (2017) 972e985.
(2019) 101e110.
~
[11] A. Krajnc, J. Brem, P. Hinchliffe, K. Calvopina, T.D. Panduwawala, P.A. Lang,
J.J.A.G. Kamps, J.M. Tyrrell, E. Widlake, B.G. Saward, T.R. Walsh, J. Spencer,
[29] K. Kwapien, M. Damergi, S. Nader, L. El Khoury, Z. Hobaika, R.G. Maroun, J.-
P. Piquemal, L. Gavara, D. Berthomieu, J.-F. Hernandez, N. Gresh, Calibration of
C.J. Schofield, Bicyclic boronate VNRX-5133 inhibits metallo- and serine
lactamases, J. Med. Chem. 62 (2019) 8544e8556.
b-
1,2,4-triazole-3-thione, an original Zn-binding group of metallo-b-lactamase
inhibitors. Validation of a polarizable MM/MD potential by quantum chem-
[12] J.C. Hamrick, J.-D. Docquier, T. Uehara, C.L. Myers, D.A. Six, C.L. Chatwin,
K.J. John, S.F. Vernacchio, S.M. Cusick, R.E.L. Trout, C. Pozzi, F. De Luca,
M. Benvenut, S. Mangani, B. Liu, R.W. Jackson, G. Moeck, L. Xerri, C.J. Burns,
D.C. Pevear, D.M. Daigle, VNRX-5133 (Taniborbactam), a broad-spectrum in-
istry, J. Phys. Chem. B 121 (2017) 6295e6312.
[30] P. Vella, W.M. Hussein, E.W. Leung, D. Clayton, D.L. Ollis, N. Mitic, G. Schenk,
ꢁ
R.P. McGeary, The identification of new metallo-b-lactamase inhibitor leads
from fragment-based screening, Bioorg. Med. Chem. Lett 21 (2011)
3282e3285.
hibitor of serine- and metallo-b-lactamase, restores activity of cefepime in
Enterobacterales and Pseudomonas aeruginosa, Antimicrob. Agents Chemo-
ther. 64 (2020) e01963-19.
[31] T. Christopeit, T.J. Carlsen, R. Helland, H.K. Leiros, Discovery of novel inhibitor
scaffolds against the metallo-b-lactamase VIM-2 by surface plasmon reso-
[13] B. Liu, R.E.L. Trout, G.H. Chu, D. McGarry, R.W. Jackson, J.C. Hamrick,
D.M. Daigle, S.M. Cusick, C. Pozzi, F. De Luca, M. Benvenuti, S. Mangani, J.-
D. Docquier, W.J. Weis, D.C. Pevear, L. Xerri, C.J. Burns, Discovery of Tani-
nance (SPR) based fragment screening, J. Med. Chem. 58 (2015) 8671e8682.
[32] F. Spyrakis, G. Celenza, F. Marcoccia, M. Santucci, S. Cross, P. Bellio, L. Cendron,
M. Perilli, D. Tondi, Structure-based virtual screening for the discovery of
novel inhibitors of New Delhi Metallo-b-lactamase-1, ACS Med. Chem. Lett. 9
(2018) 45e50.
[33] E. Freire, Do enthalpy and entropy distinguish first class from best in class?
Drug Discov. Today 13 (2008) 869e874.
borbactam (VNRX-5133): a broad spectrum serine- and metallo-b-lactamase
inhibitor for carbapenem-resistant bacterial infections, J. Med. Chem. 63
(2020) 2789e2801.
[14] M. Everett, N. Sprynski, A. Coelho, J. Castandet, M. Bayet, J. Bougnon, C. Lozano,
D.T. Davies, S. Leiris, M. Zalacain, I. Morrissey, S. Magnet, K. Holden, P. Warn,
[34] J.E. Ladbury, Calorimetry as a tool for understanding biomolecular interactions
and an aid to drug design, Biochem. Soc. Trans. 38 (2010) 888e893.
F. De Luca, J.-D. Docquier, M. Lemonnier, Discovery of a novel metallo-b-lac-
tamase inhibitor that potentiates meropenem activity against carbapenem-
resistant Enterobacteriaceae, Antimicrob. Agents Chemother. 62 (2018)
e00074-18.
ꢀ
[35] L. Borgianni, J. Vandenameele, A. Matagne, L. Bini, R. Bonomo, J.-M. Frere,
G.M. Rossolini, J.-D. Docquier, Mutational analysis of VIM-2 reveals an
essential determinant for metallo-b-lactamase stability and folding, Anti-
microb. Agents Chemother. 54 (2010) 3197e3204.
[15] S. Leiris, A. Coelho, J. Castandet, M. Bayet, C. Lozano, J. Bougnon, J. Bousquet,
M. Everett, M. Lemonnier, N. Sprynski, M. Zalacain, T.D. Pallin, M.C. Cramp,
N. Jennings, G. Raphy, M.W. Jones, R. Pattipati, B. Shankar,
R. Sivasubrahmanyam, A.K. Soodhagani, R.R. Juventhala, N. Pottabathini,
S. Pothukanuri, M. Benvenuti, C. Pozzi, S. Mangani, F. De Luca, G. Cerboni, J.-
D. Docquier, D.T. Davies, SAR studies leading to the identification of a novel
[36] C. Mugnaini, F. Sannio, A. Brizzi, R. Del Prete, T. Simone, T. Ferraro, F. De Luca,
F. Corelli, J.-D. Docquier, Screen of unfocused libraries identifies compounds
with direct or synergistic antibacterial activity, ACS Med. Chem. Lett. 11
(2020) 899e905.
[37] F.M. Studier, Protein production by auto-induction in high density shaking
cultures, Protein Expr. Purif. 41 (2005) 207e234.
series of metallo-b-lactamase inhibitors for the treatment of carbapenem-
resistant Enterobacteriaceae infections that display efficacy in an animal
infection model, ACS Infect. Dis. 5 (2019) 131e140.
[38] N. Laraki, N. Franceschini, G.M. Rossolini, P. Santucci, C. Meunier, E. de Pauw,
ꢀ
G. Amicosante, J.-M. Frere, M. Galleni, Biochemical characterization of the
[16] R.P. McGeary, D.T. Tan, G. Schenk, Progress toward inhibitors of metallo-b-
Pseudomonas aeruginosa 101/1477 metallo-b-lactamase IMP-1 produced by
Escherichia coli, Antimicrob. Agents Chemother. 43 (1999) 902e906.
lactamases, Future Med. Chem. 9 (2017) 673e691.
ꢁ
[17] B.M. Lienard, G. Garau, L. Horsfall, A.I. Karsisiotis, C. Damblon, P. Lassaux,
[39] J.-D. Docquier, F. Pantanella, F. Giuliani, M.C. Thaller, G. Amicosante,
ꢀ
C. Papamicael, G.C. Roberts, M. Galleni, O. Dideberg, J.-M. Frere, C.J. Schofield,
ꢀ
M. Galleni, J.-M. Frere, K. Bush, G.M. Rossolini, CAU-1, a subclass B3 metallo-
b
-
Structural basis for the broad-spectrum inhibition of metallo-b-lactamases by
lactamase of low substrate affinity encoded by an ortholog present in the
Caulobacter crescentus chromosome, Antimicrob. Agents Chemother. 46
(2002) 1823e1830.
thiols, Org. Biomol. Chem. 6 (2008) 2282e2294.
[18] P. Lassaux, M. Hamel, M. Gulea, H. Delbrück, P.S. Mercuri, L. Horsfall,
ꢀ
D. Dehareng, M. Kupper, J.-M. Frere, K. Hoffmann, M. Galleni, C. Bebrone,

22
L. Gavara et al. / European Journal of Medicinal Chemistry 208 (2020) 112720
ꢀ
[40] J.-D. Docquier, J. Lamotte-Brasseur, M. Galleni, G. Amicosante, J.M. Frere,
G.M. Rossolini, On functional and structural heterogeneity of VIM-type met-
counterparts, Biol. Chem. 386 (2005) 41e52.
[47] J.-D. Docquier, M. Benvenuti, V. Calderone, M. Stoczko, N. Menciassi,
G.M. Rossolini, S. Mangani, High-resolution crystal structure of the subclass B3
allo-
b
-lactamases, J. Antimicrob. Chemother. 51 (2003) 257e266.
ꢀ
[41] M. Hernandez-Villadares, M. Galleni, J.-M. Frere, A. Felici, M. Perilli,
N. Franceschini, G.M. Rossolini, A. Oratore, G. Amicosante, Overproduction and
metallo-b-lactamase BJP-1: rational basis for substrate specificity and inter-
action with sulfonamides, Antimicrob. Agents Chemother. 54 (2010)
4343e4351.
purification of the Aeromonas hydrophila CphA metallo-b-lactamse expressed
in Escherichia coli, Microb. Drug Resist. 2 (1996) 253e256.
[48] C. Pozzi, F. Di Pisa, F. De Luca, M. Benvenuti, J.-D. Docquier, S. Mangani,
[42] C. Bebrone, C. Anne, K. De Vriendt, B. Devreese, G.M. Rossolini, J. Van Beeu-
Atomic-resolution structure of a class C b-lactamase and its complex with
avibactam, ChemMedChem 13 (2018) 1437e1446.
ꢀ
men, J.;-M. Frere, M. Galleni, Dramatic broadening of the substrate profile of
the Aeromonas hydrophila CphA metallo-
genesis, J. Biol. Chem. 280 (2005) 28195e28202.
b
-lactamase by site-directed muta-
[49] G.N. Murshudov, P. Shubak, A.A. Lebedev, N.S. Pannu, R.A. Steiner,
R.A. Nicholls, M.D. Winn, F. Long, A.A. Vagin, REFMAC5 for the refinement of
macromolecular crystal structures, Acta Crystallogr. D Biol. Crystallogr. 67
(Pt4) (2011) 355e367.
[50] M.D. Wynn, C.C. Ballard, K.D. Cowtan, E.J. Dodson, P. Emsley, P.R. Evans,
R.M. Keegan, E.B. Krissinel, A.G. Leslie, A. McCoy, S.J. McNicholas,
G.N. Murshudov, N.S. Pannu, E.A. Potterton, H.R. Powell, R.J. Read, A. Vagin,
K.S. Wilson, Overview of the CCP4 suite and current developments, Acta
Crystallogr. D Biol. Crystallogr. 67 (Pt4) (2011) 235e242.
[51] P. Emsley, B. Lohkamp, W.G. Scott, K. Cowtan, Features and development of
coot, Acta Crystallogr. D Biol. Crystallogr. 66 (Pt4) (2010) 486e501.
[52] G. Langer, S.X. Cohen, V.S. Lamzin, A. Perrakis, Automated macromolecular
model building for X-ray crystallography using ARP/wARP version 7, Nat.
Protoc. 3 (2008) 1171e1179.
[43] Clinical Laboratory Standard Institute, Performance Standards for Antimicro-
bial Disk Susceptibility Tests; Approved Standard, Document M02-A12, 2015.
Twelfth Edition, Wayne, PA, USA.
[44] S. Yang, S.R. Clayton, E.L. Zechiedrich, Relative contributions of the AcrAB,
MdfA and NorE efflux pumps to quinolone resistance in Escherichia coli,
J. Antimicrob. Chemother. 51 (2003) 545e556.
[45] Clinical Laboratory Standard Institute, Methods for Dilution Antimicrobial
Susceptibility Tests for Bacteria that Grow Aerobically, Document M07-A10,
2015. Twelfth Edition, Wayne, PA, USA.
[46] M. Akoachere, R. Iozef, S. Rahlfs, M. Deponte, B. Mannervik, D.J. Creighton,
H. Schirmer, K. Becker, Characterization of the glyoxalases of the malarial
parasite Plasmodium falciparum and comparison with their human