NH-1,2,3-triazole inhibitors of the VIM-2 metallo-β-lactamase

Article abstract of DOI:10.1021/ml900022q

Metallo-β-lactamases (MBLs) are an emerging cause of bacterial resistance to antibiotic treatment. The VIM-2 β-lactamase is the most commonly encountered MBLs in clinical isolates worldwide. Described here are potent and selective small molecule inhibitor

Full text of DOI:10.1021/ml900022q

pubs.acs.org/acsmedchemlett
NH-1,2,3-Triazole Inhibitors of the VIM-2
Metallo-β-Lactamase
Timo Weide,†,^ S. Adrian Saldanha,‡,^ Dmitriy Minond,‡,^ Timothy P. Spicer,‡
§
‡
§
Joseph R. Fotsing,†, Michael Spaargaren, Jean-Marie Frere, Carine Bebrone,
ꢀ
K. Barry Sharpless,† Peter S. Hodder,*,‡ and Valery V. Fokin*,†
†Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037,
‡Lead Identification, Translational Research Institute, The Scripps Research Institute, Scripps Florida, 130 Scripps Way, Jupiter,
§
ꢀ
ꢁ
^
ꢀ
Florida 33458, and Centre for Protein Engineering, University of Liege, Allee du 6 Aout B6, Sart-Tilman 4000 Liege, Belgium
ABSTRACT Metallo-β-lactamases (MBLs) are an emerging cause of bacterial
resistance to antibiotic treatment. The VIM-2 β-lactamase is the most commonly
encountered MBLs in clinical isolates worldwide. Described here are potent and
selective small molecule inhibitors of VIM-2 containing the arylsulfonyl-NH-1,2,3-
triazole chemotype that potentiate the efficacy of the β-lactam, imipenem, in
Escherichia coli.
KEYWORDS Metallo-β-lactamases, VIM-2, NH-1,2,3-triazole-based inhibitors, bac-
terial resistance, Escherichia coli
etallo-β-lactamases(MBLs)areanemergingclassof
enzymes with the ability to degrade most β-lactam
antibacterial agents.1 As a consequence of the
were readily prepared as shown in Figure 2. The key trans-
formation was the Banert cascade reaction of propargyl
azides with various nucleophiles.11 The sequence began with
the synthesis of N-(4-chlorobut-2-ynyl)-sulfonamide deriva-
tives (b) obtained from propargyl amines (a) and sulfonyl
chlorides under the Schotten-Baumann conditions. Nucleo-
philic substitution of the chloride with sodium azide gave the
desired propargyl azides (c), which, upon heating, under-
went a facile rearrangement to triazafulvenes (d), highly
reactive intermediates that could be readily captured with
different nucleophiles.
From our initial work with arylsulfonyltriazole VIM-2
inhibitors, we noted that an alkoxy substituent on the C-4
methyl of the triazole moiety was important for activity.10
We speculate that the oxygen of the alkoxy group, a hydro-
gen-bond acceptor, is important for effective inhibition of
VIM-2, possibly through polar interactions between this
atom and the active site amino acid residues. Consequently,
to expand the SAR around the VIM-2 inhibitor 1, five new
analogues bearing an alkoxy appendage on the C-4 methyl
of the triazole moiety were synthesized (compounds 2-4, 5,
and 8 in Table 1) and examined in a nitrocefin biochemical
assay10 using VIM-2 cloned from a Pseudomonas aeruginosa
clinical strain (COL-1). Three of the five new analogues
exhibited improved potency as compared to the original
hit, 1.
M
rising rates of infections attributed to bacteria with acquired
MBL genes, enzymes of this class represent new targets for
adjuvant antibacterial therapy.2,3 The VIM-2 enzyme is
currently the most widespread MBL found in clinical isolates
worldwide,4-6 and as such, small molecule inhibitors of this
enzyme may prove useful in the treatment of resistant
infections. Several MBL inhibitors have been reported,7,8
with the most potent VIM-2 inhibitor, a 3-mercaptopropionic
acid derivative, exhibiting a Ki of 190 nM.9 However, apart
from p-chloromercuribenzoic acid (pCMB), a nonspecific
cysteine-reactive reagent, no VIM-2 inhibitor has been
shown to potentiate the antibacterial effects of β-lactams
in VIM-2-expressing bacteria.10 We recently identified a
family of 4,5-disubstituted NH-1,2,3-triazoles (Figure 1) of
selective VIM-2 inhibitors through the biochemical screening
of a focused NH-1,2,3-triazole library with 267 members.10
The most potent candidate from this study, compound 1,
N-((4-((but-3-ynyloxy)methyl)-1H-1,2,3-triazol-5-yl)methyl)-
4-iodo-benzenesulfonamide, exhibited submicromolar ac-
tivity Ki = 0.41 ( 0.03 μM against VIM-2 while being
inactive toward the related MBL, IMP-1.10 Unfortunately, this
compound did not potentiate the antibacterial effects of a
β-lactam antibiotic, imipenem.
Herein, we describe our efforts to examine structure-
activity relationships (SARs) for this chemotype with the goal
of improving the potency of the lead compound 1 and
identifying VIM-2 inhibitors with the ability to potentiate β-
lactam antibacterials. In our study, substituents on the C-4
methyl of the triazole (R1) or the arylsulfonamide (R2) were
systematically varied. NH-triazole-containing sulfonamides
Next, to further explore the substitution requirements of
the triazole moiety and to show that VIM-2 inhibition was not
Received Date: December 18, 2009
Accepted Date: March 19, 2010
Published on Web Date: April 15, 2010
r
2010 American Chemical Society
150
DOI: 10.1021/ml900022q ACS Med. Chem. Lett. 2010, 1, 150–154
|

pubs.acs.org/acsmedchemlett
unique to analogues containing a 4-iodobenzene group, a
diverse series of 4-chlorobenzenesulfonyl derivatives were
prepared and tested (Table 2). Inhibitors with submicromo-
lar IC50 values were identified from this group of candidates
(compounds 12-16); all of them contained an amine at the
C-4 methyl of the triazole ring. As for the SAR, the potency of
the 4-chlorobenzene sulfonamides was strongly dependent
on the nature of the C-4 methyl triazole substituent. In this
series, hydrophobic amino (e.g., 12, IC50 = 0.29 μM) and
alkoxy derivatives (e.g., 17, IC50 = 1.1 μM) were excellent
inhibitors, while alkyl derivatives (i.e., 25, 27 and 28) were
poorly active or inactive toward VIM-2. A sulfide 30, which is
the thioether analogue of 22, was inactive. These findings
lend further support to our hypothesis that a hydrogen-bond
acceptor of the C-4 methyl of the triazole is important for
inhibitor-protein interactions, with nitrogen being superior
to oxygen. However, with respect to VIM-2 binding and
inhibition, sulfur does not appear to be a good replacement,
probably as a consequence of its larger size and more diffuse
electronic properties, leading to poor hydrogen-bond accep-
tor properties.
A series of unsubstituted arylsulfonamides were also
examined (Table 3). Again, amino substitution on the C-4
methyl of the triazole resulted in the most potent com-
pounds, with the adamantyl (32) and cyclohexyl (33) deri-
vatives approaching IC50 = 100 nM.
Next, we looked at the effect of aromatic substituents.
The equivalent, unsubstituted 4-iodobenzenesulfonyl and 4-
chlorobenzenesulfonyl derivatives (Table 3) exhibited at
most a 3-fold difference in potency, demonstrating that
changes at the para-position provide minor improvements
in inhibitor binding.
To further explore the contribution to binding provided by
the arylsulfonamide moiety, we prepared analogues with the
same triazole derivative but with differing aromatic substit-
uents (Table 4). Again, halide aromatic substitution had
discrete effects. For example, the potency of 2, a 4-iodoben-
zenesulfonyl derivative, and 36, a 2,5-dichlorobenzenesul-
fonyl derivative, which both share an isopropyl appendage
on the triazole, differed by 2-fold. Similarly 42, a 4-methox-
ybenzenesulfonyl derivative, and 14, a 4-chlorobenzenesul-
fonyl derivative, both sharing a dipropylamine appendage
on the triazole, were equipotent. In contrast, derivatives with
very bulky aromatic para-substituents (i.e., 38-40) were
poorly active or inactive against VIM-2.
Figure 1. 4,5-Disubstituted NH-1,2,3-triazoles with substituents
on C-4 methyl of the triazole (R1) or the arylsulfonamide (R2).
Candidate 1 is shown.
During the course of lead optimization of 1, we recognized
that the most potent inhibitors contained a dichlorobenzene
group (Table 5) and an amine substituent on the C-4 methyl
of the triazole. Table 5 compares the most potent VIM-2
inhibitor 44 with other amino and alkoxy derivatives in the
Figure 2. Synthesis of candidate compounds.
Table 1. SAR of Triazolylmethyl 4-Iodobenzenesulfonamides with Respect to the C-4 Methyl Substituent
a IC50 ( error (μM) was reported as the standard deviation of three replicates. IC50 values were obtained using a four-parameter sigmoidal
dose-response curve with adjustable baseline using GraphPad Prism. Maximum inhibition (at 56 μM) was reported for compounds that do not achieve
50% inhibition. b Compounds were reported previously.10
r
2010 American Chemical Society
151
DOI: 10.1021/ml900022q ACS Med. Chem. Lett. 2010, 1, 150–154
|

pubs.acs.org/acsmedchemlett
Table 2. SAR of Triazolylmethyl 4-Chlorobenzenesulfonamides with Respect to the C-4 Methyl Substituent
Table 3. SAR of Triazolylmethyl Arylsulfonamides with Respect to the C-4 Methyl Substituent
ID
R1
IC50 (μM)
ID
R1
R2
-l
IC50 (μM)
ID
R1
R2
-I
IC50 (μM)
32
33
34
35
-NHadamantyl
-NHcyclohexyl
-NHpyrrolidine
-OiPr
0.12 ( 0.01
0.13 ( 0.01
0.19 ( 0.01
0.85 ( 0.07
2
17
3
-OiPr
-OiPr
-OtBu
-OtBu
0.29 ( 0.03
1.1 ( 0.1
1.1 ( 0.3
1.4 ( 0.1
11
22
6
-OBu
-OBu
-OMe
-OMe
5.2 ( 0.70
18.9 ( 6.00
7.3 ( 1.9
21 ( 5
-Cl
-I
-Cl
-I
19
-Cl
23
-Cl
Table 4. SAR of Triazolylmethyl Arylsulfonamides with Varying Aromatic Substituents
ID
R
IC50 (μM)
ID
R
IC50 (μM)
ID
R
IC50 (μM)
2
36
35
17
37
38
39
40
4-I
2,5-dichloro-
4-H
0.29 ( 0.03
0.61 (0.09
0.85 ( 0.07
1.10 ( 0.10
1.10 (1.00
30%
41
42
14
43
2,5-dichloro-
4-OMe
4-Cl
0.10 ( 0.01
0.33 ( 0.03
0.45 ( 0.02
1.03 ( 0.003
44
45
34
13
2,5-dichloro-
3,4-dichloro-
4-H
0.07 (0.003
0.07 (0.007
0.13 ( 0.01
0.39 ( 0.02
4-Cl
4-F
4-Cl
4-Br
4-NHAc
4-CF3
4-tBu
4.3%
0%
2,5-dichlorobenzenesulfonyl series. Compounds 44 and 47
only differ by replacement of nitrogen by oxygen and clearly
demonstrate a 20-fold improvement in potencyof the amino
over the alkoxy derivative.
To ascertain the biochemical selectivity for this chemical
class, all compounds reported here were tested against
another class B1 MBL, IMP-1. Because VIM-2 and IMP-1
show considerable sequence divergence in their active sites
r
2010 American Chemical Society
152
DOI: 10.1021/ml900022q ACS Med. Chem. Lett. 2010, 1, 150–154
|

pubs.acs.org/acsmedchemlett
Table 5. SAR of Triazolylmethyl 2,5-Dichlorobenzenesulfonamides with Respect to the C-4 Methyl Substituent
ID
R1
IC50 (μM)
ID
R1
IC50 (μM)
ID
R1
IC50 (μM)
44
46
-NHcyclohexyl
-NHadamantyl
0.07 ( 0.003
0.07 ( 0.01
41
36
-NPr2
-OiPr
0.10 ( 0.01
0.61 ( 0.09
47
-Ocyclohexyl
1.40 ( 0.50
Table 6. Results of Ki and MIC Potentiation Studies
imipenem MIC (μg/mL)
20 μMc
ID
R1
R2
Ki(μM)a
mechanismb
50 μMc
10 μMc
46
45
41
44
33
34
42
15
12
32
14
47
18
13
-NHadamantyl
-NHcyclohexyl
-NHdipropyl
-NHcyclohexyl
-NHcyclohexyl
-NHpyrrolidine
-NHdipropyl
-Nindoline
2,5-Cl
3,4-Cl
2,5-Cl
2,5-Cl
H
0.01 (0.002
0.02 ( 0.002
0.02 ( 0.004
0.03 ( 0.003
0.03 ( 0.01
0.03 ( 0.004
0.06 ( 0.01
0.06 þ 0.01
0.08 ( 0.01
0.01 þ0.001
0.12 þ 0.03
0.17 ( 0.02
0.18 þ 0.03
0.04 ( 0.01
C
C
C
C
C
C
M
C
C
C
C
C
C
C
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
0.617
1.851
0.617
1.851
1.851
0.617
0.617
1.851
0.617
1.851
1.851
1.851
1.851
1.851
1.851
1.851
1.851
1.851
1.851
1.851
1.851
H
methoxy
Cl
-NHPh
Cl
-NHadamantyl
-NHdipropyl
-Ocyclohexyl
-NHt butylamino
-NHcyclohexyl
H
Cl
2,5-Cl
Cl
Cl
a Ki values were calculated by nonlinear regression (hyperbolic equation) using GraphPad Prism. b C = competitive inhibition, and M = mixed
inhibition. c Inhibitor concentration used in the potentiation assay.
(33% overall sequence identity), the arylsulfonyltriazoles
were not expected to inhibit both enzymes unless behaving
through promiscuous mechanisms. Indeed, all 47 com-
pounds were inactive against IMP-1.
absence of a carbapenem, imipenem. Relative to the par-
ental (BL21) strain, the MIC (minimum inhibitory con-
centration) for imipenem increased approximately 9-fold
when cells were transformed with the VIM-2-encoding plas-
mid (imipenem MIC = 0.21 μg/mL in BL21 vs imipenem
MIC = 1.85 μg/mL in BL21/VIM-2). This decrease in antibac-
terial potency reflects the ability of VIM-2 to degrade imipe-
nem. When tested at 50 μM, 14 of the 47 arylsulfonyltriazoles
potentiated imipenem and consequently were tested at lower
doses (Tabl e 6). Compound 45, with the best potentiation,
improved the MIC of imipenem by 3-fold from 1.85 to 0.617
μg/mL at 10 μM (Tabl e 6). To ascertain whether this class of
inhibitor can fully potentiate imipenem, compound 45 was
tested at higher concentrations. Indeed, the MIC for imipenem
was fully restored to that observed for the parental strain (i.e.,
lacking VIM-2) when 45 was tested at 150 μM. Furthermore, in
parental (BL21) cells, the MIC for imipenem was unaffected by
the presence of 45 as high as 150 μM. It was also established
that none of these 14 compounds exhibited intrinsic antibac-
terial activity since E. coli cell growth was unaffected when
they were tested at 150 μM.
Docking studies with 4412 suggest that this inhibitor binds
to VIM-2 in the same mode as was predicted for the parent, 1,
which is described elsewhere.10 SAR studies show that the
best inhibitors (i.e., 41 and 44-46) display hydrophobic
groups at the C-4 methyl substituent of the triazole. Our
docking and modeling studies have identified a cavity that
can accommodate a compact hydrophobic group such as
adamantyl or cyclohexyl.10,12 Consistent with SAR findings,
smaller triazole appendages cannot take full advantage of this
cavity, while groups larger or more extended than adamantyl
would be detrimental to binding due to steric clashes.
Currently, only pCMB, a slowly reversible/irreversible VIM-
2 inhibitor, has been shown to have a synergistic effect with
β-lactam antibacterials in VIM-2-expressing bacteria.10 How-
ever, this cysteine-reactive reagent is known to have several
off-target activities, lessening its value for mechanistic stu-
dies. To demonstrate potentiation by the arylsulfonyltria-
zoles in bacteria, we assessed the ability of these compounds
to affect the growth of parental (BL21) and VIM-2-trans-
formed (BL21/VIM-2) Escherichia coli, in the presence or
The mechanism of inhibition for the 14 VIM-2 inhibitors
active at 50 μM in VIM-2-expressing E. coli was further eval-
uated through kinetic studies (Table 6). The best inhibitors
r
2010 American Chemical Society
153
DOI: 10.1021/ml900022q ACS Med. Chem. Lett. 2010, 1, 150–154
|

pubs.acs.org/acsmedchemlett
exhibited Ki values as low as 10 nM (i.e., 46), and all but one
exhibited classical competitive inhibition and hence bind at
the active site. Only 42 exhibits a mixed mode of inhibition
with features of both competitive and noncompetitive in-
hibition. While inhibition kinetics for 42 are somewhat
different as compared to the other 13 compounds, a com-
petitive inhibition component still suggests that binding of
42 occurs at the active site.
In conclusion, the fidelity and robustness of the Banert
cascade have been used to rapidly assemble an NH-1,2,3-
triazole library from diverse arrays of highly functionalized
fragments. Medicinal chemistry efforts progressed via three
iterative steps and resulted in the synthesis of 320 unique
compounds. The starting compound exhibited moderate
biochemical potency toward VIM-2 while being inactive
toward IMP-1. Substitution at the arylsulfonamide produced
subtle improvements in potency. On the other hand, ana-
logues bearing amino substitution on the C-4 methyl of the
triazole generated highly potent VIM-2 inhibitors that po-
tentiate imipenem antibacterial activity. The best inhibitors
exhibit as much as 40-fold increase in potency in bioche-
mical assays over 1 and represent the most potent VIM-2
inhibitors to date. Furthermore, these are the first reported
inhibitors to be active against VIM-2 in bacteria (E. coli), with
no apparent off-target effects.
(2)
Cornaglia, G.; Akova, M.; Amicosante, G.; Canton, R.; Cauda,
R.; Docquier, J. D.; Edelstein, M.; Frere, J. M.; Fuzi, M.; Galleni,
M.; Giamarellou, H.; Gniadkowski, M.; Koncan, R.; Libisch, B.;
Luzzaro, F.; Miriagou, V.; Navarro, F.; Nordmann, P.; Pagani,
L.; Peixe, L.; Poirel, L.; Souli, M.; Tacconelli, E.; Vatopoulos, A.;
Rossolini, G. M. Metallo-beta-lactamases as emerging resistance
determinants in Gram-negative pathogens: Open issues. Int.
J. Antimicrob. Agents 2007, 29, 380–388.
(3)
(4)
Livermore, D. M. The threat from the pink corner. Ann. Med.
2003, 35, 226–234.
Bebrone, C. Metallo-beta-lactamases (classification, activity,
genetic organization, structure, zinc coordination) and their
superfamily. Biochem. Pharmacol. 2007, 74, 1686–1701.
Walsh, T. R.; Toleman, M. A.; Poirel, L.; Nordmann, P. Metallo-
beta-lactamases: The quiet before the storm? Clin. Microbiol.
Rev. 2005, 18, 306–325.
Docquier, J. D.; Lamotte-Brasseur, J.; Galleni, M.; Amicosante,
G.; Frere, J. M.; Rossolini, G. M. On functional and structural
heterogeneity of VIM-type metallo-beta-lactamases. J. Anti-
microb. Chemother. 2003, 51, 257–266.
Lienard, M. A.; Strandh, M.; Hedenstrom, E.; Johansson, T.;
Lofstedt, C. Key biosynthetic gene subfamily recruited for
pheromone production prior to the extensive radiation of
Lepidoptera. BMC Evol. Biol. 2008, 8, 270.
Olsen, L.; Jost, S.; Adolph, H. W.; Pettersson, I.; Hemmingsen,
L.; Jorgensen, F. S. New leads of metallo-beta-lactamase
inhibitors from structure-based pharmacophore design.
Bioorg. Med. Chem. 2006, 14, 2627–2635.
(5)
(6)
(7)
(8)
(9)
Jin, W.; Arakawa, Y.; Yasuzawa, H.; Taki, T.; Hashiguchi, R.;
Mitsutani, K.; Shoga, A.; Yamaguchi, Y.; Kurosaki, H.; Shibata,
N.; Ohta, M.; Goto, M. Comparative study of the inhibition of
metallo-beta-lactamases (IMP-1 and VIM-2) by thiol com-
pounds that contain a hydrophobic group. Biol. Pharm. Bull.
2004, 27, 851–856.
SUPPORTING INFORMATION AVAILABLE Method of en-
zyme inhibition, microbial inhibition, docking studies, and syn-
thetic procedures and characterization. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
(10) Minond, D.; Saldanha, S. A.; Subramaniam, P.; Spaargaren,
M.; Spicer, T.; Fotsing, J. R.; Weide, T.; Fokin, V. V.; Sharpless,
Corresponding Author: *To whom correspondence should be
addressed. (P.S.H.) Tel: (561) 228-2100. Fax: (+1 ) 561-228-3054.
E-mail: hodderp@scripps.edu. (V.V.F) Tel: 858-784-7515. Fax: 858-
784-7562. E-mail: fokin@scripps.edu.
ꢀ
K. B.; Galleni, M.; Frere, J.-M.; Hodder, P. Inhibitors of VIM-2
by Screening Pharmacogically Active and Click-Chemistry
Compound Libraries. Bioorg. Med. Chem. 2009, 15, 5027–
5037.
(11) Loren, J. C.; Sharpless, K. B. The banert cascade: A synthetic
sequence to polyfunctional NH-1,2,3-triazoles. Synthesis-
Stuttgart 2005, 1514–1520.
Present Addresses: Current address: Senomyx, Inc., 4767
Nexus Center Dr., San Diego, CA 92121.
(12) See the Supporting Information for docking studies with
compound 44.
Author Contributions: ^ These authors made equal contribu-
tions to this work.
Funding Sources: The National Institutes of Health (NS059451,
P.S.H.; GM087620, V.V.F.) and the Skaggs Institute for Chemical
Biology (K.B.S. and V.V.F.) supported this work.
ACKNOWLEDGMENT Pierre Baillargeon and Lina DeLuca are
thanked for their assistance with compound management and
Louis Scampavia for LC-MS analysis of the presented compounds
(Lead Identification Division, Translational Research Institute,
Scripps, Florida). Moreno Galleni (d'Enzymologie & Centre d'In-
genierie des Proteines, Institut de Chimie) is thanked for helpful
discussions.
REFERENCES
(1)
Jacoby, G. A.; Munoz-Price, L. S. The new beta-lactamases.
N. Engl. J. Med. 2005, 352, 380–391.
r
2010 American Chemical Society
154
DOI: 10.1021/ml900022q ACS Med. Chem. Lett. 2010, 1, 150–154
|