Chromophore-linked substrate (CLS405): Probing metallo-β-lactamase activity and inhibition

Article abstract of DOI:10.1002/cmdc.201300350

Serine- and metallo-β-lactamases present a threat to the clinical use of nearly all β-lactam antibiotics, including penicillins, cephalosporins, and carbapenems. Efforts to develop metallo-β-lactamase (MBL) inhibitors require suitable screening platforms to allow the rapid determination of β-lactamase activity and efficient inhibition. Unfortunately, the platforms currently available are not ideal for this purpose. Further progress in MBL inhibitor identification requires inexpensive and widely applicable assays. Herein the identification of an inexpensive and stable chromogenic substrate suitable for use in assays of clinically relevant MBLs is described. (6R,7R)-3-((4-Nitrophenoxy)methyl)-8-oxo-7-(2-phenylacetamido) -5-thia-1-azabicyclo[4.2.0]oct-2-ene-2-carboxylic acid 5,5-dioxide (CLS405) was synthesised in a three-step protocol. CLS405 was then characterised spectroscopically, and its stability and kinetic properties evaluated. With a Δλ_max value of 100 nm between the parent and hydrolysis product, a higher analytical accuracy is possible with CLS405 than with commonly used chromogenic substrates. The use of CLS405 in assays was validated by MBL activity measurements and inhibitor screening that resulted in the identification of N-hydroxythiazoles as new inhibitor scaffolds for MBLs. Further evaluation of the identified N-hydroxythiazoles against a panel of clinically relevant MBLs showed that they possess inhibitory activities in the mid- to low-micromolar range. The findings of this study provide both a useful tool compound for further inhibitor identification, and novel scaffolds for the design of improved MBL inhibitors with potential as antibiotics against resistant strains of bacteria. Monitoring MBLs! Resistance to β-lactam antibiotics, mediated by metallo-β-lactamases (MBLs), is an increasing clinical problem. While compounds that target MBLs could be useful antibacterial agents, their identification is hampered by the lack of suitable assay platforms. To this end, CLS405, a chromophore-linked MBL substrate, was developed and its applicability demonstrated by the identification of N-hydroxythiazoles as potential inhibitors against a panel of clinically relevant MBLs. Copyright

Full text of DOI:10.1002/cmdc.201300350

CHEMMEDCHEM
FULL PAPERS
DOI: 10.1002/cmdc.201300350
Chromophore-Linked Substrate (CLS405): Probing Metallo-
b-Lactamase Activity and Inhibition
[a]
[a]
[a]
[b, c]
Anne Makena, Sander S. van Berkel, Clarisse Lejeune, Raymond J. Owens,
[
b, c]
[d]
[d]
[a]
Anil Verma,
Ramya Salimraj, James Spencer, Jꢀrgen Brem,* and
[a]
Christopher J. Schofield*
Serine- and metallo-b-lactamases present a threat to the clini-
cal use of nearly all b-lactam antibiotics, including penicillins,
cephalosporins, and carbapenems. Efforts to develop metallo-
b-lactamase (MBL) inhibitors require suitable screening plat-
forms to allow the rapid determination of b-lactamase activity
and efficient inhibition. Unfortunately, the platforms currently
available are not ideal for this purpose. Further progress in
MBL inhibitor identification requires inexpensive and widely
applicable assays. Herein the identification of an inexpensive
and stable chromogenic substrate suitable for use in assays of
clinically relevant MBLs is described. (6R,7R)-3-((4-Nitrophenox-
y)methyl)-8-oxo-7-(2-phenylacetamido)-5-thia-1-azabicyclo-
ised spectroscopically, and its stability and kinetic properties
evaluated. With a Dlmax value of 100 nm between the parent
and hydrolysis product, a higher analytical accuracy is possible
with CLS405 than with commonly used chromogenic sub-
strates. The use of CLS405 in assays was validated by MBL ac-
tivity measurements and inhibitor screening that resulted in
the identification of N-hydroxythiazoles as new inhibitor scaf-
folds for MBLs. Further evaluation of the identified N-hydroxy-
thiazoles against a panel of clinically relevant MBLs showed
that they possess inhibitory activities in the mid- to low-micro-
molar range. The findings of this study provide both a useful
tool compound for further inhibitor identification, and novel
scaffolds for the design of improved MBL inhibitors with po-
tential as antibiotics against resistant strains of bacteria.
[4.2.0]oct-2-ene-2-carboxylic acid 5,5-dioxide (CLS405) was syn-
thesised in a three-step protocol. CLS405 was then character-
Introduction
b-Lactam antibiotics have been in use for more than half a cen-
with increased industrialisation and mobility, has resulted in
[3]
tury, and remain the most used group of antibiotics worldwide
the global spread of a plethora of b-lactamases. b-Lactamases
can be classified into two major groups: the serine-b-lactamas-
es (SBLs), which use a nucleophilic serine residue in the active
site for catalysis, and the metallo-b-lactamases (MBLs), which
contain one or two zinc ions in their active site that facilitate
[
1]
(
over 60%). They are antibiotics of choice because of their af-
fordability, oral bioavailability, high clinical efficacy, and low
[
2]
toxicity. However, the efficacy of b-lactam antibiotics is de-
clining due to bacterial resistance mechanisms including the
widespread occurrence of b-lactamases, which catalyze the hy-
drolysis of the b-lactam ring, rendering them inactive. The in-
troduction of new generations of b-lactam antibiotics, coupled
[4]
catalysis. SBLs are further categorised as Ambler classes A
[5]
(penicillinases), C (cephalosporinases), and D (oxacillinases).
Mechanism-based inhibitors of SBLs such as clavulanic acid, ta-
zobactam, and sulbactam have been identified and are of
+
+
[6,7]
[
a] A. Makena, Dr. S. S. van Berkel, Dr. C. Lejeune, Dr. J. Brem,
widespread clinical utility.
Prof. Dr. C. J. Schofield
Department of Chemistry, University of Oxford
The MBLs (Ambler class B) catalyze the hydrolysis of virtually
all b-lactams except monobactams and are not susceptible to
any of the clinically used serine-b-lactamase inhibitors. MBLs
can be further classified into three subclasses (B1, B2, and B3)
1
2 Mansfield Rd, Oxford, OX1 3TA (UK)
E-mail: jurgen.brem@chem.ox.ac.uk
christopher.schofield@chem.ox.ac.uk
[8]
[b] Dr. R. J. Owens, Dr. A. Verma
according to their sequence homology. MBLs were long con-
sidered to be of academic, rather than clinical, interest. Howev-
er, in recent years the widespread emergence of MBLs in clini-
cally important bacterial pathogens has changed this situation.
To date, the most clinically relevant identified MBLs are from
the B1 class (e.g., from the NDM, IMP and VIM families). This is
due to their ability to hydrolyze “last resort” antibiotics, such
as recent generations of carbapenems and cephalosporins cou-
pled with their spread among opportunist Gram-negative bac-
terial pathogens (both Enterobacteriaceae and non-fermenters
such as Pseudomonas aeruginosa) by plasmid/transposon-medi-
Oxford Protein Production Facility UK, University of Oxford
Research Complex at Harwell, Rutherford Appleton Laboratory
Harwell Oxford, Didcot OX11 0FA (UK)
[c] Dr. R. J. Owens, Dr. A. Verma
Division of Structural Biology
Henry Wellcome Building for Genomic Medicine
Roosevelt Drive, Oxford OX3 7BN (UK)
[d] R. Salimraj, Dr. J. Spencer
School of Cellular & Molecular Medicine, University of Bristol
Medical Sciences Building, University Walk, Bristol BS8 1TD (UK)
+
[
] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201300350.
[9]
ated horizontal gene transfer. Previous studies have identified
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1923 – 1929 1923

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
various types of MBL inhibitors including biphenyl tetrazoles,
homocysteinyl peptides, mercaptocarboxylates, and succinic
chromacef, or a leaving group (e.g., CENTA) which when ex-
pelled allows for the detection of a UV-active chromophore.
However, the currently used cephalosporin-based chromogenic
substrates suffer from drawbacks including low stability, light
sensitivity, and high cost (in particular for nitrocefin), which
limit their application in HTS. In pursuit of an inexpensive,
stable, and widely applicable chromogenic substrate for MBLs,
we describe the synthesis and biological evaluation of a novel
cephalosporin-based chromogenic substrate for the applica-
tion in MBL assays including inhibitor screening. The chromo-
phore-linked substrate (CLS405) was tested against five clinical-
ly relevant MBLs: IMP-1 (Imipenemase), NDM-1 (New Delhi
MBL), SPM-1 (S¼o Paulo MBL), VIM-1 and VIM-2 (Verona imipe-
nemase), and the “model” MBL Bc-II (Bacillus cereus II). The ki-
netic properties of CLS405 for the various MBLs were evaluat-
ed and compared with the commonly used chromogenic sub-
strates CENTA and nitrocefin. The use of CLS405 as a reporter
substrate was exemplified by screening of a small set of poten-
tial MBL inhibitors leading to the identification of hydroxythia-
zoles as a new type of MBL inhibitor.
[
10–12]
acid derivatives.
However, a clinically useful inhibitor of
a panel of MBLs has not yet been reported. Such an MBL inhib-
itor has the potential to extend the lifetime of important b-
lactam antibiotics as has been the case for the three marketed
SBL inhibitors (i.e., clavulanic acid, tazobactam, and sulbactam).
The rapid global dissemination of MBLs coupled to the lack of
a clinically useful inhibitor is an increasing public health con-
[
13]
cern.
The discovery of potent, broad-spectrum MBL inhibitors is
limited in part by the development of expedient assay plat-
forms for high-throughput screening (HTS). b-Lactamase assays
are also widely used as reporters for gene expression studies.
The majority of b-lactamase assays are based on the ability of
the enzyme to catalyze the hydrolysis of the b-lactam ring
leading to a distinct change in UV/Vis absorbance between the
[
14]
substrate and the resultant hydrolyzed product. Other colori-
metric-based b-lactamase assays include an iodometric-based
method that depends on the reduction of iodine by the hydro-
lysed product and subsequent decolourisation of the starch–
[
15]
iodine complex. Another method is based on the change in
Results and Discussion
pH following b-lactam hydrolysis which uses colorimetric pH
[
16]
indicators such as phenol red or bromocresol. These meth-
ods, however, have been superseded by the use of more
robust and versatile chromogenic and fluorogenic substrates.
Commonly used chromogenic substrates include cephalospor-
In an attempt to identify a broadly applicable and cost-efficient
chromogenic substrate for MBLs suitable for inhibitor screen-
ing, we evaluated various chromophore-based cephalosporins.
Given that para-nitrophenol has been frequently used in pro-
[
17]
[18]
[19]
[21]
in-based compounds such as PADAC, CENTA, nitrocefin,
tease and esterase screens
as a chromogenic leaving
[
20]
[22,23]
and chromacef (Figure 1).
group,
we investigated whether nitrophenol-linked cepha-
Typically, cephalosporin-based chromogenic substrates have
a chromophore attached at the C3 position; which can either
be a highly conjugated group, as in the case of nitrocefin and
losporins could be used in MBL assays. This led to the identifi-
cation of CLS405 as a candidate substrate which releases
a para-nitrophenolate (pNP) anion upon b-lactam hydrolysis.
Synthesis
The target chromogenic sub-
strate CLS405 was produced
using a straightforward three-
step protocol starting from com-
mercially available para-methox-
ybenzyl ether (PMB)-protected
chlorocephalosporin (Scheme 1).
The first step involved the con-
version of a protected C3’ chlor-
ocephalosporin into the corre-
sponding C3’ iodocephalosporin,
followed by a subsequent substi-
tution reaction with para-nitro-
phenol. The resulting product
was obtained in reasonable yield
(
57%) as a mixture of D -D iso-
2 3
mers 1. This mixture was treated
with two equivalents of meta-
chloroperbenzoic acid (mCPBA)
to produce the sulfone 2 as
a
single isolated D3 isomer
Figure 1. a) Structures of nitrocefin, CENTA, chromacef, and CLS405; b) metallo-b-lactamase (MBL)-triggered hy-
drolysis of substrate CLS405 and subsequent release of the detectable para-nitrophenolate (pNP) anion.
(52%) (note that the use of
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1923 – 1929 1924

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
quence similarities between the various MBLs and is
consistent with differences in catalytic efficiency for
a given substrate, as well as the inhibitor susceptibili-
ty, that have been previously observed. In all cases,
the turnover number (k ) was found to be highest
cat
for CLS405 among the substrates tested. For IMP-
3
ꢀ1
1
the turnover rate was as high as 54.6ꢂ10 s . The
relatively high kcat values for CLS405 may, in part, re-
flect a favorable positioning of the sulfone moiety in
the active site, possibly facilitating the hydrolysis of
the substrate. For all enzymes tested, the k /K
cat
m
value was substantially higher for CLS405 than for
the other tested substrates. In the case of NDM-1,
the k /K for CLS405 was 70-fold higher than for the
Scheme 1. The three-step protocol for the synthesis of chromogenic substrate CLS405.
cat
m
Reagents and conditions: a) NaI (10 equiv), acetone, RT, 2 h; b) K
phenol (1 equiv), RT, 2.5 h (57% over 2 steps); c) meta-chloroperbenzoic acid (2 equiv),
CH Cl , 08C, 30 min, then RT, overnight (52%); d) CF CO H/anisole (10:1), 08C, 30 min
52%).
2
CO
3
, MeCN, para-nitro-
other tested chromogenic substrates. Taken together,
these data indicate that CLS405 is a superior sub-
strate for MBL inhibitor screening (including for the
2
2
3
2
(
1
equivalent of mCPBA under the conditions applied did not
result in the formation of the expected sulfoxide, instead the
sulfone 2 and some starting material were present as indicated
by TLC analysis). The final step involved acid-mediated depro-
tection of the PMB-protecting group (CF CO H/anisole 10:1) to
Table 1. Kinetic parameters for the three chromogenic substrates tested
on six metallo-b-lactamases (MBLs).
[
a]
[b]
Enzyme
NDM-1
Substrate
[E]
[pm]
k
[s
cat
K
[mm]
m
k
[mm
cat/K
m
ꢀ1
ꢀ1 ꢀ1
3
2
]
s
]
efficiently produce the desired chromogenic substrate CLS405
52%).
Nitrocefin
CENTA
CLS405
1000
1000
50
25.3
92.5
2116
8.8ꢁ1
34.6ꢁ7
10.2ꢁ1
2.9
2.7
(
208.5
Spectroscopic characterisation
Nitrocefin
CENTA
CLS405
50
100
10
2794
431.9
54.6ꢂ10
55.7ꢁ5
17.1ꢁ2
73.9ꢁ6
50.2
25.3
738.5
IMP-1
VIM-1
VIM-2
SPM-1
Bc-II
3
The maximum absorbance wavelength of CLS405 was found
to be 305 nm, which differs significantly from that of the C3’
leaving group that is liberated on hydrolysis (i.e., pNP, labs
Nitrocefin
CENTA
CLS405
500
1000
250
130.1
133.6
435.2
151.1ꢁ1
98.4ꢁ11
30.1ꢁ2
0.9
1.4
14.5
=
405 nm, see Figure 2). As a consequence, hydrolysis of CLS405
manifests a Dlmax of 100 nm compared with a Dlmax of 59 nm
for CENTA and its C3’ leaving group (5-mercapto-2-nitrobenzo-
ic acid) as generated upon hydrolysis. As a result of the large
Dlmax of CLS405, the peak overlap is significantly less than for
the other substrates, thus enabling higher analytical accuracy
in monitoring of either substrate degradation or product for-
mation.
Nitrocefin
CENTA
CLS405
100
1000
40
225.6
48.2
1254.5
7.2ꢁ1
26.1ꢁ3
21.6ꢁ2
31.2
1.8
58.0
3
3
Nitrocefin
CENTA
CLS405
50ꢂ10
10ꢂ10
250
0.7
6.1
414.8
16.1ꢁ4
25.3ꢁ6
10.9ꢁ1
0.04
0.2
38.2
Nitrocefin
CENTA
CLS405
1000
1000
400
14.9
8.9
206.0
8.1ꢁ1
136ꢁ16
64.8ꢁ6
1.8
0.06
3.2
Kinetic evaluation
Comparison of the sensitivity of the various chromogenic sub-
strates revealed that substrate CLS405 is highly responsive, al-
lowing the use of extremely low MBL concentrations (mostly in
the picomolar range, Table 1), with IMP-1 requiring no more
than 10 pm enzyme. The enzyme concentration for NDM-1 was
up to 20-fold lower with CLS405 than for the other tested
chromogenic substrates (Table 1). The highest enzyme concen-
tration used was 400 pm for Bc-II, which was still less than half
the enzyme concentration used for the other substrates.
[
a] Values for kcat, K
m
and kcat/K
m
for NDM-1, IMP-1, VIM-2, SPM-1 and Bc-II
with Nitrocefin and CENTA were taken from Ref. [24]. [b] Standard devia-
tion values for kcat did not exceed 10%.
widely disseminated NDM-1 enzyme) relative to the other com-
monly used chromogenic substrates.
Monitoring the stability of CLS405, we found that a solution
of CLS405 could be used for 1–2 days when stored on ice.
Moreover, DMSO stock solutions, stored at ꢀ208C, could be
used up to four months after preparation without any detecta-
ble decomposition. Storage of CLS405 as a solid at 48C al-
lowed the use of this compound over a prolonged period of
time (months). Monitoring the rate of CLS405 hydrolysis over
a pH range of 3–10 (see the Supporting Information) indicated
When comparing the K values of CLS405, CENTA, and nitro-
m
cefin for the different MBLs, no clear trend was observed. Km
values between 10 and 70 mm were derived for CLS405 for the
different MBLs tested, thus enabling subsequent inhibitor
screens to be carried out with <100 mm substrate concentra-
tion. These variations in K values reflect the relatively low se-
m
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1923 – 1929 1925

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
was initially screened at a single concentration (100 mm) in trip-
licate, and as a cutoff parameter compounds that yielded a re-
sidual activity of <5% was selected for further investigation.
Comparison of the residual activities of the various classes of
metalloenzyme inhibitors, as shown in Table 2, indicated that
the majority of the tested succinic acids (3, 4b, 4c), hydroxya-
mic acids (5a–c) and thiols (6c–e) had residual activities
>
20%. On the other hand, the dithiol 6a, the hydroxythiazole
ethyl esters (7a, 7c, 7e), and the carboxylic acid derivative 7b
gave residual activities of <5%. These compounds, including
compound 6a, were subsequently selected for IC50 determina-
tions.
Dithiol 6a was found to be the most potent inhibitor for all
the enzymes tested, with IC50 values ranging from 0.5 to
3.1 mm (Table 3). Compound 6a was previously identified as
a very potent inhibitor of the monozinc MBL CphA, and
a good inhibitor of the IMP-1 and Bc-II enzymes, with K values
i
[29]
<
1 mm. The results presented herein suggest that deriva-
tives of 6a merit further investigation as broad-spectrum MBL
inhibitors.
Hydroxythiazoles have been reported as potential inhibitors
of ferrous iron and 2-oxoglutarate (2OG)-dependent oxygenas-
es that are involved in hypoxia-dependent transcriptional regu-
[28,30]
lation.
We found that this compound class, that is, the
esters (7a, 7c, 7e) and the corresponding carboxylic acid
forms (7b, 7 f), displayed substantial activity against one or
more MBLs. (Note that we cannot exclude partial ester hydroly-
sis, either enzyme- or non-enzyme-mediated, under the stan-
dard incubation conditions.) The IC50 values for the N-hydroxy-
thiazole ethyl ester 7a showed good activity against NDM-
1
1
(9.1ꢁ0.7 mm) but no significant inhibition of SPM-1 and IMP-
, which have been reported to have similar substrate profi-
[31]
les. The corresponding carboxylic acid of 7a, that is, 7b, also
showed a good inhibitory effect against NDM-1 (2.5ꢁ0.6 mm)
and was a good inhibitor for the other MBLs tested except
VIM-2, and to a lesser extent Bc-II (Table 3). The N-hydroxythia-
zole 7c also showed good activity against NDM-1 (6.1ꢁ
0
.6 mm) and SPM-1 (2.8ꢁ1.1 mm), while still maintaining IC
5
0
values of <20 mm for the other MBLs.
The corresponding carboxylic acid 7d, however, showed
a high residual activity (59%) and thus was not considered for
IC50 determination. Compound 7e showed a relatively low IC50
value for NDM-1 (12.6ꢁ0.9 mm), while its corresponding acid
Figure 2. Substrate (S, c) and product (P, a) absorption spectra for
a) CENTA (S: 347 nm; P: 406 nm), b) nitrocefin (S: 389 nm; P: 496 nm), and
CLS405 (S: 305 nm; P: 405 nm). See the Supporting Information for experi-
mental details.
7
f (control) gave a higher IC50 value (26.7ꢁ9.2 mm), corre-
sponding with the RA values observed for these compounds.
Interestingly, compound 7 f showed good activity against SPM-
1 (4.5ꢁ1.2 mm). Overall, the results indicate emergent struc-
ture–activity relationships within the small set of hydroxythia-
zoles tested. The differences in inhibition profiles of the tested
compounds exemplify the challenge of obtaining a broad-
spectrum MBL inhibitor that is presented by the structural and
sequence diversity of MBLs.
that CLS405 is stable across a pH range of 6–8, with >85% of
the intact substrate remaining after 12 h incubation at room
temperature.
Inhibitor screening
To test the suitability of CLS405-based assays for inhibitor
screening we tested four classes of previously reported
The “cross-reactivity” of the hydroxythiazoles as inhibitors of
[28]
both MBLs and oxygenases illustrates an advantage of cross-
screening, but also suggests that cross-family selectivity stud-
ies may be relevant in the characterisation of both zinc-depen-
dent MBL and iron-dependent oxygenase inhibitors.
[
25]
metalloenzyme inhibitors comprising thiols,
hydroxyamic
[
26]
[27]
[28]
acids, succinic acid analogues, and N-hydroxythiazoles
as potential MBL inhibitors. A small set of these compounds
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1923 – 1929 1926

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
chromogenic substrates (CLS405, CENTA, and nitro-
cefin) and six MBLs (NDM-1, VIM-1, VIM-2, IMP-1,
SPM-1, and Bc-II), revealed substantial differences in
the kinetic parameters for different enzyme–sub-
strate combinations. The chromogenic substrate
CLS405 was found to be a highly sensitive and effi-
cient substrate for the five tested clinically relevant
[
a]
Table 2. Residual activities (RA) of different types of metalloenzyme inhibitors.
Compd
RA [%]
Compd
RA [%]
65
3
20
6c
ꢀ
1
ꢀ1
MBLs, having k /Km values of >15 mm s . This
cat
4
4
a
<5
6d
6e
33
66
substrate was successfully applied in an exemplary
screening assay for the identification of MBL inhibi-
tors. Several N-hydroxythiazoles displayed inhibitory
activity against our panel of MBLs, with compound
b
40
7
c showing inhibition with IC50 values <20 mm
against all the tested clinically relevant MBLs. Like-
wise, 7b gave IC50 values of <10 mm for all the clini-
cally relevant enzymes except VIM-2. The results
suggest that, in addition to their use as inhibitors of
iron(II)-dependent oxygenases, hydroxythiazoles
may have utility as inhibitors of zinc(II)-dependent
hydrolases, and specifically MBLs.
4
5
5
5
6
6
c
a
b
c
100
83
7a <5
7b <5
7c <5
Experimental Section
78
Protein production
For the production of recombinant Bc-II, IMP-1, and
NDM-1 MBLs, previously reported plasmids pET-9a Bc-
75
7d
59
[
32]
[33]
II, pET-26b IMP-1, and pTriEx-based pOPIN-F NDM-
1
[34]
were used. For the production of VIM-1, VIM-2, and
SPM-1 MBLs, pOPIN-F plasmids based on the pTriEx
[35]
a
b
<5
<5
7e <5
vector were used. These vectors use E. coli optimised
codons (Genscript, Piscataway, NJ 08854, USA). These
plasmids were used to produce full-length VIM-2 (27–
2
66) and SPM-1 (29–276) with a cleavable 3C protease
7 f
36
N-terminal (His) tag; full-length VIM-1 (2–266) was pro-
6
duced with a C-terminal (His) tag.
6
The recombinant MBLs were produced in E. coli
BL21(DE3) pLysS cells using 2TY media supplemented
with the appropriate antibiotic. For Bc-II and IMP-
[
a] Residual activities were determined at an inhibitor concentration of 100 mm using
NDM-1 (50 pm) and a final substrate concentration of 10 mm. Napth=naphthalene;
Ph=phenyl; Et=ethyl.
ꢀ
1
1
both kanamycin (50 mgmL ) and chloramphenicol
ꢀ1
(
34 mgmL ) were used, while for
VIM-1, VIM-2, NDM-1, and SPM-
combination of ampicillin
[
a]
Table 3. Inhibitory activities of six potential broad-spectrum metallo-b-lactamase (MBL) inhibitors.
1
(
a
ꢀ1
50 mgmL ) and chloramphenicol
Enzyme
IC50 [mm]
ꢀ1
6
a
7a
7b
7c
7e
7 f
(34 mgmL ) was used. Cells were
grown at 378C until the OD600
NDM-1
IMP-1
VIM-1
VIM-2
SPM-1
Bc-II
1.9ꢁ0.9
0.5ꢁ0.3
ND
3.1ꢁ1.2
0.7ꢁ0.3
2.4ꢁ1.4
9.1ꢁ0.7
2.5ꢁ0.6
3.5ꢁ0.9
6.1ꢁ0.6
18.6ꢁ7.7
16.2ꢁ3.3
11.4ꢁ3.3
2.8ꢁ1.1
8.0ꢁ2.2
12.6ꢁ0.9
26.7ꢁ9.2
154.8ꢁ46.2
19.3ꢁ9.6
81.6ꢁ19.4
4.5ꢁ1.2
reached 0.6–0.7 before induction
with 0.5 mm IPTG. The cells were
grown at the same temperature
for 4 h for Bc-II, IMP-1, and VIM-1.
The temperature was then lowered
to 208C (for NDM-1) and 308C (for
>500
>500
61.6ꢁ21.1
27.5ꢁ6.8
6.2ꢁ1.3
31.5ꢁ5.5
40.3ꢁ16.9
73.6ꢁ19.7
20.5ꢁ8.2
161.5ꢁ20.1
2.2ꢁ0.8
>500
20.7ꢁ13.5
80.9ꢁ80.1
60.8ꢁ20.4
[
a] Measurements were carried out in triplicate, and IC50 values are the average ꢁ standard error of the mean;
VIM-2 and SPM-1), and the cells
ND: not determined.
were further incubated for 20 h for
NDM-1, 8 h for SPM-1, and 4 h for
VIM-2. Cells were harvested by centrifugation (10 g, 10 min) and re-
suspended in 50 mL lysis buffer supplemented with DNase, lyso-
syme, and EDTA-free protease inhibitor. The cells were further
lysed by sonication (2ꢂ7 min, 10-s pulse at 60% amplitude with
10 s rest times), and cell debris removed by centrifugation (23 g,
30 min).
Conclusions
We have developed a stable and inexpensive chromogenic de-
tection substrate (CLS405) suitable for the identification of
novel MBL inhibitors. A comparative study, employing three
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1923 – 1929 1927

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
In the case of VIM-1, VIM-2, NDM-1 and SPM-1, cell lysates were
loaded onto a 5 mL HisTrap HP column (GE Healthcare Life Scien-
ces, Little Chalfont, UK), washed with 50 mm Tris pH 7.5, 500 mm
NaCl, containing 20 mm imidazole and eluted with an imidazole
gradient (up to 500 mm imidazole). The proteins were further puri-
fied using a Superdex S200 column (300 mL) and eluted with
sion of Cl-Cep-OPMB ester (257.5 mg, 0.5 mmol) in acetone (3 mL)
was treated with NaI (752 mg, 5.0 mmol, 10 equiv). The reaction
was stirred at RT for 2 h, after which the solvent was removed in
vacuo. The crude mixture was dissolved in EtOAc (20 mL) and
washed with H O (3ꢂ10 mL), 5% aq Na S O (2ꢂ20 mL) and brine
2
2
2
3
(20 mL). The organic layer was dried over Na SO , filtered, and con-
2
4
2
0 mm Tris pH 7.5, 200 mm NaCl. All the proteins except VIM-1,
centrated in vacuo. The crude product was redissolved in MeCN
(5 mL), and the solution was treated with para-nitrophenol
(139 mg, 1.0 mmol, 2 equiv) and K2CO3 (207 mg, 1.5 mmol,
3 equiv). The reaction was monitored by TLC (c-Hexane/EtOAc,
1:1), and completion was reached after 2.5 h. The solvent was
evaporated in vacuo, and the resulting crude mixture was parti-
which had a C-terminal (His) tag, were incubated overnight at 48C
6
with 3C protease (1:100 w/w) and then purified using a 5 mL Hi-
sTrap HP column to give the untagged protein. In the case of Bc-II
and IMP-1, cell lysates were dialyzed against 20 mm MES pH 6.4
supplemented with 0.2 mm ZnCl2 for Bc-II or 50 mm Tris pH 7.0
supplemented with 0.1 mm ZnCl2 for IMP-1 and loaded onto an
SP-Sepharose column (1.5ꢂ12 cm with a 25 mL bed volume),
which was also pre-equilibrated with same buffer. Bound proteins
were eluted with a 0!1m NaCl gradient in the appropriate buffer.
IMP-1 was further purified by loading onto a Superdex S75 column
tioned between H O (10 mL) and EtOAc (10 mL). After separation
2
of the layers, the aqueous layer was extracted with EtOAc (2ꢂ
15 mL), and the combined organic layers were washed with 5% aq
Na S O (2ꢂ20 mL) and brine (20 mL). The organic layer was dried
2
2
3
over Na SO , filtered, and concentrated in vacuo. The crude prod-
2
4
(
1
300 mL) and eluting with 50 mm Tris buffer (pH 7.5) containing
50 mm NaCl and 100 mm ZnCl2.
uct was purified by column chromatography (c-Hexane/EtOAc, 1:1)
to yield 1 as a pale-yellow solid (169 mg, 57%): R =0.35 (c-
F
Hexane/EtOAc, 1:1). The ratio of D /D isomers was determined to
be approximately 4:1 according to H NMR analysis.
3
1
2
For all the MBLs, the purity of the resulting fractions was ascer-
tained to be >90% by SDS-PAGE analysis (see the Supporting In-
formation), and fractions containing purified protein were concen-
trated by centrifugal ultrafiltration. The concentrations of the puri-
fied proteins were determined using a ND-1000 NanoDrop spectro-
photometer and the consecutive concentrations used in the assay
were calculated based on the original concentration.
1
D -isomer (major): H NMR (400 MHz, CDCl ): d=8.13 (d, J=9.5 Hz,
3
3
2
6
8
5
4
H), 7.28–7.41 (m, 5H), 7.16 (d, J=8.5 Hz, 2H), 6.85–6.89 (m, 1H),
.75–6.83 (m, 4H), 6.41 (s, 1H), 6.14 (d, J=9.0 Hz, 1H), 5.67 (dd, J=
.5, 4.0 Hz, 1H), 5.27 (d, J=4.0 Hz, 1H), 5.16 (d, J=11.5 Hz, 1H),
.11 (d, J=1.5 Hz, 1H), 5.02 (d, J=11.5 Hz, 1H), 4.56–4.61 (m, 1H),
.44–4.48 (m, 1H), 3.80 (s, 3H), 3.66 ppm (d, J=4.0 Hz, 2H);
13
C NMR (101 MHz, CDCl ): d=171.0, 166.8, 164.2, 162.8, 160.1,
3
1
41.9, 133.5, 130.4 (2C), 129.5 (2C), 129.2 (2C), 127.8, 126.3, 125.8
Substrate synthesis
(2C), 121.4, 118.8, 114.6 (2C), 114.1 (2C), 70.1, 68.0, 60.5, 55.3, 53.6,
Chemicals were from commercial suppliers (Aldrich, Acros, Alfa
Aesar, and TCI) and were used without further purification. (6R,7R)-
ꢀ
4
9.8, 43.3 ppm; LRMS (ESIꢀ): m/z (%): 588.0 [MꢀH] (90%).
4
-Methoxybenzyl 3-(chloromethyl)-8-oxo-7-(2-phenylacetamido)-5-
(6R,7R)-4-Methoxybenzyl 3-((4-nitrophenoxy)methyl)-8-oxo-7-(2-
phenylacetamido)-5-thia-1-aza-bicyclo[4.2.0]oct-2-ene-2-carbox-
ylate 5,5-dioxide (2): A suspension of 1 (165 mg, 0.28 mmol) in
thia-1-azabicyclo[4.2.0]oct-2-ene-2-carboxylate (para-methoxyben-
zyl (PMB)-protected Cep-Cl) was obtained from Activate Scientific
(
Prien, Germany). Solvents (including dry solvents) for chemical
dry CH Cl (10 mL) was cooled to 08C and treated with mCPBA
2 2
transformations, work-up and chromatography were HPLC grade
from Aldrich (Dorset, UK) and used without prior distillation. Silica
gel 60 F254 analytical thin layer chromatography (TLC) plates were
purchased from Merck (Darmstadt, Germany) and visualized under
(77% pure, 125 mg, 0.56 mmol, 2 equiv). The reaction was stirred
at 08C for 30 min (formation of sulfoxide was not observed), after
which time the reaction was warmed to RT and stirred overnight.
Upon completion, the crude reaction mixture was evaporated to
dryness, and the resultant white solid was re-suspended in cold
UV light or with KMnO stain. Chromatographic purifications were
4
performed using Merck Geduran 60 silica (40–63 mm) or prepacked
SNAP columns on a Biotage SP1 purification system (Uppsala,
Sweden).
Et
traces of mCPBA and mCBA. The fine solid was then filtered and
washed with Et O (2ꢂ5 mL) and dried under high vacuum to yield
as a white solid (92 mg, 52%): R =0.85 (CH Cl /MeOH, 9:1);
2
O (5 mL). The suspension was stirred for 2 h at RT to remove
2
2
F
2
2
Deuterated solvents were purchased from Sigma or Apollo Scientif-
1
H NMR (400 MHz, [D ]DMSO): d=8.94 (d, J=8.5 Hz, 1H), 8.19 (d,
1
13
6
ic Ltd. All H and C NMR spectra were recorded using a Bruker
J=9.5 Hz, 2H), 7.18–7.37 (m, 7H), 7.06 (d, J=9.5 Hz, 2H), 6.86 (d,
J=9.0 Hz, 2H), 6.02 (dd, J=8.5, 5.0 Hz, 1H), 5.43 (d, J=4.5 Hz, 1H),
Avance 400 MHz spectrometer. All chemical shifts (d) are given in
[36]
parts per million (ppm) relative to the residual solvent peak, and
5
3
1
.22 (s, 2H), 4.90 (s, 2H), 4.20–4.52 (m, app. AB-system, 2H), 3.71 (s,
coupling constants (J) are reported in Hertz (Hz) to the nearest
13
H), 3.53–3.65 ppm (m, 2H); C NMR (101 MHz, [D ]DMSO): d=
6
0
.5 Hz. Low-resolution mass spectrometry (LRMS) was performed
70.9, 164.5, 162.8, 160.5, 159.4, 141.3, 135.6, 130.5 (2C), 129.2
on a Waters LCT Premier instrument using an electrospray ionisa-
tion (ESI) source and Time of Flight (ToF) analyser.
(2C), 128.2 (2C), 126.6, 126.5, 125.8 (2C), 124.5, 124.1, 115.2 (2C),
1
13.8 (2C), 67.0, 66.8, 66.3, 58.4, 55.0, 50.8, 41.2 ppm; LRMS (ESIꢀ):
ꢀ
m/z (%): 620.2 [MꢀH] (100%).
For HPLC separation, the reaction mixture was dissolved in MeOH
and analysed using reverse-phase LCMS, performed on a WATERS
Sunfire C18 column (150 mmꢂ4.6 mm, 5 mm). Separation was ach-
ieved using a linear gradient of solvent A (H O+0.01% CF CO H)
(6R,7R)-3-((4-Nitrophenoxy)methyl)-8-oxo-7-(2-phenylacetami-
do)-5-thia-1-azabicyclo[4.2.0]oct-2-ene-2-carboxylic acid 5,5-di-
oxide (CLS405): Neat compound 2 (62 mg, 0.1 mmol) was cooled
2
3
2
and solvent B (MeCN+0.01% CF CO H), eluting at a flow rate of
3
2
ꢀ1
to 08C, which after a mixture of CF CO H and anisole (10:1, 1 mL)
3 2
1
mLmin and monitoring at 254 nm: 0% B over 2 min, 0!100%
was added, and the reaction was stirred at 08C for 30 min. After
B over 16 min, and 100% B over 2 min.
completion, Et O (5 mL) was added causing an off-white solid to
2
(
6R,7R)-4-Methoxybenzyl 3-((4-nitrophenoxy)methyl)-8-oxo-7-(2-
phenylacetamido)-5-thia-1-aza-bicyclo[4.2.0]oct-2-ene-2 (D3-
isomer) and oct-3-ene-2 (D2-isomer) carboxylate (1): A suspen-
form. The solid was isolated by centrifugation (4000 rpm for
4 min), washed with cold Et O (2ꢂ2 mL), and drying under high
2
vacuum to give CLS405 as an off-white solid (26 mg, 52%): R =
F
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1923 – 1929 1928

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
1
0
.10 (CH Cl /MeOH, 9:1); H NMR (400 MHz, CD OD): d=8.21–8.26
[2] M. Colomer-Lluch, J. Jofre, M. Muniesa, PLoS One 2011, 6, e17549.
2
2
3
[
[
3] P. M. Hawkey, A. M. Jones, J. Antimicrob. Chemother. 2009, 64, i3–i10.
4] D. Girlich, L. Poirel, P. Nordmann, J. Antimicrob. Chemother. 2012, 67,
(
m, 2H), 7.27–7.33 (m, 4H), 7.24 (m, 1H), 7.12–7.19 (m, 2H), 5.98 (d,
J=4.5 Hz, 1H), 5.22 (d, J=5.0 Hz, 1H), 5.11–5.18 (m, app. q, 2H),
2661–2664.
4
.35–4.07 (m, app. AB-system, 2H), 3.57–3.69 ppm (m, app. q, 2H);
1
3
[5] R. P. Ambler, Philos. Trans. R. Soc. London Ser. B 1980, 289, 321–331.
C NMR (101 MHz, [D ]DMSO): d=170.9, 164.3, 163.0, 162.1, 141.3,
6
[
[
[
6] R. A. Bonomo, Clin. Infect. Dis. 2011, 52, 485–487.
7] A. Matagne, M. F. Ghuysen, J.-M. Frꢃre, Biochem. J. 1993, 295, 705–711.
8] C. Bebrone, Biochem. Pharmacol. 2007, 74, 1686–1701.
1
6
35.6, 129.2 (2C), 128.2 (2C), 126.5, 125.92 (2C), 125.87, 115.3 (2C),
6.8, 66.5, 64.9, 58.3, 50.8, 41.2 ppm; LRMS (ESIꢀ): m/z (%): 524.0
ꢀ
[M+Na] (70%).
[9] H. C. Maltezou, Int. J. Antimicrob. Agents 2009, 33, 405.e1-405.e7.
10] J. Spencer, T. R. Walsh, Angew. Chem. 2006, 118, 1038–1042; Angew.
Chem. Int. Ed. 2006, 45, 1022–1026.
11] Faridoon, N. Ul Islam, Sci. Pharm. 2013, 81, 309–327.
12] W. Fast, L. D. Sutton, Biochim. Biophys. Acta Proteins Proteomics 2013,
[
The synthesis of N-hydroxythiazole-based inhibitors 7b, 7d, 7 f,
h, 7j was performed according to the procedure of Tegley
7
[
[
[
28]
et al. For details, see the Supporting Information.
1
834, 1648–1659.
[
13] T. R. Walsh, M. A. Toleman, J. Antimicrob. Chemother. 2012, 67, 1–3.
Kinetic analyses
[14] T. Viswanatha, L. Marrone, V. Goodfellow, G. I. Dmitrienko, Methods Mol.
Med. 2008, 142, 239–260.
To investigate the suitable enzyme concentrations for kinetic meas-
urements, the change in absorbance at 405 nm for CLS405 was
monitored as function of varying enzyme concentrations (12.5 pm
to 1 nm). Steady-state kinetic studies were carried out using BMG
Labtech Pheraster FS and NovaStar microplate readers. Hydrolysis
of nitrocefin or CENTA was monitored by determining the changes
in absorbance at l 492 and 405 nm, respectively. Kinetic constants
[
15] T. Sawai, I. Takahashi, S. Yamagishi, Antimicrob. Agents Chemother. 1978,
3, 910–913.
1
[16] T. J. Lucas, J. Clin. Pathol. 1979, 32, 1061–1065.
[17] R. N. Jones, H. W. Wilson, W. J. Novick, Jr., J. Clin. Microbiol. 1982, 15,
677–683.
[
18] C. Bebrone, C. Moali, F. Mahy, S. Rival, J. D. Docquier, G. M. Rossolini, J.
Fastrez, R. F. Pratt, J.-M. Frꢃre, M. Galleni, Antimicrob. Agents Chemother.
2
001, 45, 1868–1871.
(
Km and k ) were obtained by determining the initial rate of hy-
cat
[
19] C. H. O’Callaghan, A. Morris, M. S. Kirby, A. H. Shingler, Antimicrob.
Agents Chemother. 1972, 1, 283–288.
20] S. Yu, A. Vosbeek, K. Corbella, J. Severson, J. Schesser, L. D. Sutton, Anal.
Biochem. 2012, 428, 96–98.
drolysis (i.e., product formation) at different substrate concentra-
tions. Various enzyme concentrations were used for the panel of
MBLs studied (Table 1) with the substrate concentrations ranging
between 0.4 and 500 mm. The Michaelis–Menten curves were fitted
and analyzed using GraphPad Prism 5.01 software.
[
[21] M. Hriscu, L. Chi s¸ , M. To s¸ a, F. D. Irimie, Eur. J. Lipid Sci. Technol. 2013,
115, 571–575.
[
22] M. Adam, C. Damblon, B. Plaitin, L. Christiaens, J.-M. Frꢃre, Biochem. J.
990, 270, 525–529.
[23] J. C. Moore, F. H. Arnold, Nat. Biotechnol. 1996, 14, 458–467.
1
Inhibitor screening
[
24] S. S. van Berkel, J. Brem, A. M. Rydzik, R. Salimraj, R. Cain, A. Verma, R. J.
Owens, C. W. G. Fishwick, J. Spencer, C. J. Schofield, J. Med. Chem. 2013,
56, 6945–6953.
Inhibitors from an in-house collection were screened for NDM-1 in-
hibitory activity using CLS405 as the reporter substrate. The MBL
concentrations used were the same as those used for kinetic analy-
ses, and the concentration of the substrate was at or near the Km
value. The rate of hydrolysis of CLS405 (10 mm) by NDM-1 (50 pm)
was determined in the presence of 100 mm solutions of the inhibi-
tors. Subsequently, IC50 values were determined for three com-
pounds with a residual activity of <5%. IC50 determinations were
carried out by pre-incubating the MBLs and inhibitors at 258C for
[25] S. Siemann, A. J. Clarke, T. Viswanatha, G. I. Dmitrienko, Biochemistry
2003, 42, 1673–1683.
[
26] B. M. R. Liꢄnard, L. E. Horsfall, M. Galleni, J.-M. Frꢃre, C. J. Schofield,
Bioorg. Med. Chem. Lett. 2007, 17, 964–968.
[
27] J. H. Toney, G. G. Hammond, P. M. D. Fitzgerald, N. Sharma, J. M. Balko-
vec, G. P. Rouen, S. H. Olson, M. L. Hammond, M. L. Greenlee, Y. D. Gao,
J. Biol. Chem. 2001, 276, 31913–31918.
[
28] C. M. Tegley, V. N. Viswanadhan, K. Biswas, M. J. Frohn, T. A. N. Peterkin,
C. Chang, R. W. Bꢀrli, J. H. Dao, H. Veith, N. Rogers, S. C. Yoder, G. Biddle-
come, P. Tagari, J. R. Allen, R. W. Hungate, Bioorg. Med. Chem. Lett. 2008,
18, 3925–3928.
1
0 min in the assay buffer, prior to the addition of the substrate, as
[24]
has been done in previous MBL inhibition studies. The residual
activities were obtained at increasing inhibitor concentrations (0.8–
4
00 mm), and the data were analyzed with GraphPad Prism 5.01
[29] B. M. Liꢄnard, G. Garau, L. Horsfall, A. I. Karsisiotis, C. Damblon, P. Las-
saux, C. Papamicael, G. C. Roberts, M. Galleni, O. Dideberg, J.-M. Frꢃre,
C. J. Schofield, Org. Biomol. Chem. 2008, 6, 2282–2294.
software. The compounds were similarly tested against the panel
of remaining MBLs.
[
30] I. K. H. Leung, M. Demetriades, A. P. Hardy, C. Lejeune, T. J. Smart, A.
Szçllçssi, A. Kawamura, C. J. Schofield, T. D. W. Claridge, J. Med. Chem.
2
013, 56, 547–555.
Acknowledgements
[
31] T. A. Murphy, A. M. Simm, M. A. Toleman, R. N. Jones, T. R. Walsh, Antimi-
crob. Agents Chemother. 2003, 47, 582–587.
This research was supported by the Rhodes Trust (UK), the Well-
come Trust (UK), the Biotechnology & Biological Sciences Re-
search Council (BBSRC), the Medical Research Council (MRC)/
Canadian grant G1100135, and the Cancer Research UK (CRUK).
The Oxford Protein Production Facility UK (OPPF-UK) is supported
by the MRC and BBSRC.
[32] D. de Seny, C. Prosperi-Meys, C. Bebrone, G. M. Rossolini, M. I. Page, P.
Noel, J.-M. Frꢃre, M. Galleni, Biochem. J. 2002, 363, 687–696.
[
[
[
[
33] D. H. Griffin, T. K. Richmond, C. Sanchez, A. J. Moller, R. M. Breece, D. L.
Tierney, B. Bennett, M. W. Crowder, Biochemistry 2011, 50, 9125–9134.
34] V. L. Green, A. Verma, R. J. Owens, S. E. Phillips, S. B. Carr, Acta Crystal-
logr. Sect. F 2011, 67, 1160–1164.
35] N. S. Berrow, D. Alderton, S. Sainsbury, J. Nettleship, R. Assenberg, N.
Rahman, D. I. Stuart, R. J. Owens, Nucleic Acids Res. 2007, 35, e45.
36] H. E. Gottlieb, V. Kotlyar, N. Abraham, J. Org. Chem. 1997, 62, 7512–
7515.
Keywords: antibiotics · assays · chromogenic substrates · drug
resistance · metallo-b-lactamases · pan inhibitors
Received: August 29, 2013
[1] R. P. Elander, Appl. Microbiol. Biotechnol. 2003, 61, 385–392.
Published online on October 25, 2013
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1923 – 1929 1929